Where Are We Going Now? The Current and Future Distributions of the Monk Parakeet (Myiopsitta monachus) and Eurasian Collared Dove (Streptopelia decaocto) in a Megalopolis

Abstract

The monk parakeet (Myiopsitta monachus) and the Eurasian collared dove (Streptopelia decaocto) are two of the most prevalent invasive species globally due to their high dispersal ability. Since these birds were first recorded (1999 for the monk parakeet and 2013 for Eurasian collared dove) in the Mexico City Metropolitan Area (MCMA), both species have spread rapidly throughout the area. However, the impacts of global climate changes on the distribution patterns of these species remain poorly studied across the MCMA. Therefore, based on an ecological niche modeling approach, we assessed the expansion and potential invasion of both species in this megalopolis using current and future climate projections (year 2050). Our results estimated that the current suitable areas are 5564 km² for the monk parakeet and 5489 km² for the Eurasian collared dove, covering ~70% of the study area, suggesting a rapidly invading species, as expected. We observed a slight decrease (up to 24%) in both species in future climate scenarios, but our models estimated that the sizes of the suitable areas would remain stable. We found that the range expansion of these species in the megalopolis may be largely attributed to their propensity for jump dispersion and short-time niche expansion ability. Our findings allow for a better understanding of the factors contributing to the range expansion of the monk parakeet and the Eurasian collared dove in Mexico and can better inform the monitoring guidelines for and assessments of these invasive species.

1. Introduction

Alien species are ones that have been introduced by humans into areas outside their original distribution range. Given that the introduced species represent a great threat to global biodiversity [1], only surpassed by climate change and the fragmentation of natural habitats [2,3], the rapid spread of alien species is considered an important component of the global environmental crisis. In addition, their impacts on biodiversity are exceedingly difficult to reverse due to the extent and magnitude they can reach [4]. After their introduction and successful establishment in new environments, alien species become naturalized, managing to expand their populations and participate in biological invasion [5].

Due to their rapid spread and impact after their introduction and successful establishment, the monk parakeet (Myiopsitta monachus) and the Eurasian collared dove (Streptopelia decaocto) are two recent and remarkable cases of biological invasion. These two species have been frequently released into urban and rural environments due to the pet trade, one of the main introduction pathways of invasive alien species [6,7]. Both species appear to have ecological and morphological traits that make them successful invaders across many regions of the world [8,9,10,11,12,13], including rapid adaptability to urban and suburban environments; high growth and reproduction rates; dietary flexibility; and high plasticity and phenotypic flexibility. The biological invasion of M. monachus and S. decaocto into new environments can bring negative ecological impacts, such as competition with native birds for food resources, breeding sites, and nesting materials, leading to the displacement and even local extinction of resident species [8,14,15]. Moreover, invasive species are also a potential vector for diseases that may increase the mortality of native birds and affect human health [15,16,17,18].

The monk parakeet is native to the southern Neotropics in Paraguay, Argentina, Uruguay, Brazil, and Bolivia [19,20], while the Eurasian collared dove is native to Pakistan, India, Sri Lanka, Myanmar, and Bangladesh in southern Asia [21,22]. Myiopsitta monachus and S. decaocto are also ecologically well-distributed, as both are established in different environments throughout the country [23,24,25]. The occurrences of both birds as alien species have been documented in several temperate countries, such as the United States, Spain, and France [9,10,26,27,28,29], where established populations and local expansion dynamics have been reported [30,31,32,33,34]. Both species have experienced the largest population increases recorded among exotic birds in the world, which have allowed them to establish new populations in large cities [10,18,26,35]. For example, M. monachus has become established in several high-latitude, extra-tropical cities, such as New York, Chicago, Portland, and Seattle, USA, and in British Columbia, Canada [24], which has implications for high-altitude areas such as the Mexico City Metropolitan Area.

Since the first record of these invasive birds in Mexico (M. monachus in 1999 and S. decaocto in 2000), worryingly, both species have increased their geographic ranges to more than 95 cities throughout the country [23,24,25]. Regarding their recent introduction into the country, especially in one of the most important megalopolises, Mexico City, M. monachus was first recorded in the northern part in 1999, while S. decaocto was first recorded in the northeastern part in 2013 (Figure 1). One consequence of the expansion of both species is the homogenization of biotic communities [36,37] because invasive species can decrease local biodiversity [38,39,40,41]; such a reduction in local diversity can in turn reduce the resilience of communities to environmental disturbances due to the loss of significant functional traits [42]. In addition, the impacts of global climate changes on the distribution patterns of M. monachus and S. decaocto currently remain poorly studied in urban contexts.

The development of environmental niche modeling tools has facilitated the study of ecological determinants and historical limits [43,44]. Such an approach integrates the study of climatic tolerances through combining species occurrence data and environmental variables over the species’ past, present, and future geographic range [44,45]. This approach utilizes species-specific attributes and spatially explicit climatic attributes to predict the potential spread of the target species in time and space [46]. The results derived from predicted distribution changes can then be used as inputs to objective management plans to control and prevent the expansion of naturalized species [47,48].

Several factors, such as abiotic differences, biotic interactions, dispersal barriers, and human activities, can promote range expansion during invasion processes [49]. This, in turn, can facilitate the colonization of an area with novel combinations of environmental conditions, leading a species to explore the limits of its native niche and to occupy different niches in the newly invaded range [50]. Both invasive alien species in this study have been reported to mostly occupy large urban landscapes from which they later expand into surrounding agricultural areas and other sites highly modified via human activity [33,35,51]. In this study, based on a species distribution modeling approach for these two species, and on factors such as climatic conditions driving species distribution at geographically larger scales [52,53], we assessed the factors underlying the distribution expansion of M. monachus and S. decaocto in high-elevation regions in central Mexico (specifically in the Mexico City Metropolitan Area and adjacent areas), and we identified areas with the highest potential invasion risk for these two species under future climatic scenarios (i.e., the year 2050). Our results may be used to inform efforts to eradicate or control both species, and to mitigate the possible impacts on the native biota and natural ecosystems in this region.

2. Materials and Methods

2.1. Study Area

The Mexico City Metropolitan Area (MCMA) is localized in the Trans-Mexican Volcanic Belt in central Mexico (19°25 N, 99°07 W) (Figure 1). The lowest elevation is 2240 m but increases from the center towards the high mountains (>3900 in the Ajusco volcano). Covering more than 7954 km², the MCMA is considered one of the largest megalopolises in the world [54] and encompasses large urban areas north of Mexico City and neighboring Mexico State. The rapid growth of the population in the MCMA (more than 22 million, >2559.8 hab/km²) [55,56] has led to the loss and fast transformation of the original habitats, creating a highly homogeneous system with different land use types in which only 34% of the natural vegetation remains [57,58]. Grasslands and desert scrub are the dominant vegetation in the north and east; the south-eastern lowlands are dominated by agricultural fields, and oak pine fir forests are present in relatively large areas in the south, east, and west.

2.2. Occurrence Records for Species

We compiled occurrence records for both M. monachus and S. decaocto from invaded regions across Mexico. These records were obtained from different scientific databases, such as the Atlas of the Birds of Mexico [59], the Christmas Bird Counts (available at https://netapp.audubon.org/cbcobservation/, accessed on 22 January 2024), ORNISnet (available at http://ornisnet.org/, accessed on 22 January 2024), GBIF (Global Biodiversity Information Facility) [60,61], and GAVIA (Global Avian Invasive Atlas) [62], and complemented with citizen databases, such as eBird (available at https://ebird.org/, accessed on 22 January 2024) [24,25] and iNaturalistMX (available at https://mexico.inaturalist.org/, accessed on 22 January 2024). We supplemented these data with information from field surveys performed by JER-A (from January 2015 to December 2023) throughout the MCMA and adjacent areas in Mexico State. We initially obtained 3624 occurrence records for M. monachus and 5766 for S. decaocto.

To clean the databases for model building, duplicate records and those lacking exact geocoordinate data were eliminated. We further reduced the effects of sampling bias using spThin [63] in RStudio [64]. This allowed us to estimate the most appropriate distance among the occurrence data for each species and the nearest neighbor, retaining only those records separated by at least 3 km. Likewise, to identify imprecise species occurrences with incorrect climate values [65], we applied an environmental heterogeneity filter to the occurrence data as implemented in SDMtoolbox in ArcMap 10.2 [66]. After these procedures, we obtained 144 unique occurrence records for M. monachus (from 1999 to 2023; Figure 1A) and 278 for S. decaocto (from 2013 to 2023; Figure 1B) across the MCMA and adjacent areas.

2.3. Current and Future Climatic Scenarios

To characterize Grinnellian niches (see [67,68]) for each species, we used interpolated climate data (~1 km² cell size resolution) from CHELSA 1.2 (available at https://chelsa-climate.org/, accessed on 22 January 2024). This bioclimatic dataset includes 19 climatic variables summarizing aspects of precipitation and temperature based on the interpolation of data recorded from worldwide meteorological stations spanning a period approximately covering 1980–2010 [69]. Because reducing the multidimensionality and collinearity problems of the variables is a critical part of the modeling process, we selected seven environmental variables to build the potential distribution models for both species (

Table 1. Contribution of the bioclimatic variables to the MaxEnt modeling (%) for Myiopsitta monachus and Streptopelia decaocto.

Species	Variable	Description	Percent Contribution (%)	Permutation Importance (%)
Myiopsitta monachus	Bio01	Annual mean temperature	0.4	3.2
	Bio02	Mean diurnal range	4.0	4.0
	Bio03	Isothermality	52.1	0.4
	Bio05	Max temperature of warmest month	14.6	22.8
	Bio06	Min temperature of coldest month	1.3	15.1
	Bio11	Mean temperature of coldest quarter	17.4	44.0
	Bio12	Annual precipitation	10.2	10.6
Streptopelia decaocto	Bio02	Mean diurnal range	35.4	24.5
	Bio03	Isothermality	8.6	5.2
	Bio06	Min temperature of coldest month	17.5	14.3
	Bio10	Mean temperature of warmest quarter	10.5	8.1
	Bio12	Annual precipitation	14.5	21.1
	Bio14	Precipitation of driest month	5.9	10.8
	Bio15	Precipitation seasonality	7.6	16.0

); to accomplish this, we used a Pearson’s coefficient (r < 0.8) and Variance Inflation Factor [70,71,72]. The selected environmental predictors have been previously regarded as important limiting factors controlling the ecophysiology and distribution of M. monachus and S. decaocto [73,74].

The variables for future climate projections (year 2050) were downloaded from the Intergovernmental Panel on Climate Change (IPCC) Sixth Assessment Report [75]. These scenarios are an important integrating element of the IPCC for understanding the possible climate outcomes, impacts and risks, and mitigation futures [76]. For future climate model simulations, we selected two different climate projections from General Circulations Models (GCMs) GFDL-ESM4 and IPSL-CM6A-LR; both GCMs were used based on three Shared Socio-economic Pathway (SSP) scenarios (SSPs 126, 370 and 585), which assume a high level of greenhouse gas emissions, a medium to high end of plausible future pathways, and a low incidence of climate change mitigation policies, respectively [75,77]. These two models have been identified as more appropriate and useful for ENMs [78,79,80].

2.4. Ecological Niche and Species Distribution Models

We modeled potential climatic suitability areas for both species using MaxEnt v.3.4.1., given its high performance and suitability for presence-only data [81]. The MaxEnt algorithm calculates the maximum entropy probability distribution of suitable conditions for a given species based on a set of climatic data associated with its occurrence data [81,82,83]. To achieve this, we also performed a calibration assessing model complexity [84].

The models were generated using the kuenm R package [72], which enable detailed model calibration and selection (i.e., selecting the best modeling parameters in each case), and the creation of a final model for each species based on each model´s statistical significance, performance, and simplicity. In this manner, we assessed 62 candidate solutions (n = 31 for each species), including a combination of four different sets of environmental predictors, 17 regularization multiplier values [RM: 0.5–8.0], and 11 possible combinations of linear, quadratic, product, and threshold feature types. After this, we selected the best set of candidate solutions showing both a low omission error (<0.05) and low Akaike information criterion (AICc) values [84,85,86,87]. The statistical performance of each model was evaluated using the omission error values [88], the area under the curve (AUC) [89,90], the partial ROC test [91], and TSS (True Skill Statistics) [92].

The final models were transferred to future environmental conditions [87,93,94]. All models were run allowing “unconstrained extrapolation” and “extrapolation by clamping” in MaxEnt projections, which refer to different approaches in handling predictions outside the range of environmental conditions observed in the training data [81,84]. “Unconstrained extrapolation” allows model predictions to extend beyond the range of environmental conditions present in the training data (i.e., areas where environmental conditions might be significantly different from those observed in the training data), while “extrapolation by clamping” restricts predictions to conditions similar to those observed in the training data, minimizing the risk of unrealistic predictions in completely novel environments for each species in future scenarios [81,84]. To provide further reliability of our models when depicting future projections, we performed a multivariate environmental similarity surface (MESS) analysis [95] to help determine areas with higher degrees of uncertainty (high extrapolation values) in the models, offering a view of the novelty of future climate conditions relative to present-day conditions within the calibration area [96].

Then, we visualized the habitat suitability distribution for both species in ArcGIS 10.2 [66]. Based on the continuous suitability score (0–1), we classified the habitat suitability as follows: unsuitable habitat (<0.30), low suitability habitat (0.30 to 0.60), medium-suitability habitat (0.60 to 0.80), and high-suitability habitat (>0.80) [97]. Finally, we generated potential habitat suitability maps, in which the continuous probability maps from 0 to 1 (the logistic response outputs from MaxEnt) [82] were converted into a presence-absence map by setting a decision threshold equal to a 10% omission of the training occurrence data (a 5% false negative rate). This threshold allowed a reduction in commission errors (the false-positive rate) in our final binary maps, resulting in more conservative estimates of species distributions [98].

2.5. Risk of Expansion

To quantify changes in suitable environmental conditions for M. monachus and S. decaocto under climate change, we used SDMtoolbox [65] in ArcGIS 10.2 [66] to compare habitats predicted from current climate conditions with those corresponding to future climate conditions (SSPs 126, 370 and 585). Depending on the differences between current and future scenarios, we divided the results into three categories: gain (an increase in the species’ distribution area), stable (the distribution area was not reduced or shifted), and loss (the distribution area was reduced). In addition, to facilitate the identification of geographic spaces and potential invasion areas, we overlaid the potential distributions for each species with the agricultural areas that were obtained from INEGI (available at https://www.inegi.org.mx/, accessed on 6 February 2024). We also assessed the risk of invasion in protected areas (PAs) using the inventory of the National Commission of Protected Natural Areas (CONANP, available at https://www.gob.mx/conanp, accessed on 6 February 2024).

Finally, to evaluate the expansion of the two invasive species at the regional level, we regressed the annual distances from the first detection of M. monachus and S. decaocto [23,24,25] in the MCMA as a function of time [99]. Annual distances were calculated from the first reported occurrence; from that date onwards, all other occurrences were taken into account individually under the following criteria: only occurrences at a single site were considered, taking into account the oldest occurrence, and data from subsequent occurrences at the same site were discarded; several detections in the same year were considered, as long as they were from sites other than the first occurrence or the oldest detection at the same site. Furthermore, we ran a piecewise linear regression analysis [100,101] in RStudio [64] both to identify significant breaking points and to distinguish different invasion phases. The significance of the breaking points was tested using a Davies test [102]. The slope of each segment provided an estimated rate of spread (V, measured in km/year).

3. Results

3.1. Ecological Models and Geographical Distribution Patterns

Since the first record of occurrence of each species in the MCMA, more than 3600 points of occurrence of M. monachus and more than 5700 of S. decaocto were recorded throughout the metropolitan area through 2023 (Figure 1A,B), with the greatest number of records between 2019 and 2023, representing more than 60% of the total records. Furthermore, 185 and 153 of these records for M. monachus and S. decaocto, respectively, were found in agricultural areas and protected areas.

We detected good model performance under the selected parameter settings for both M. monachus and S. decaocto (Table S1). According to the contribution rate of environmental variables (Table 1), the dominant variables for each species are the isothermality and mean temperature of the coldest quarter for M. monachus, and the mean diurnal range and minimum temperature of the coldest month for S. decaocto, representing a cumulative contribution rate higher than 50%. From the permutation importance of bioclimatic variables (Table 1), the dominant variables for each species are the mean temperature of the coldest quarter and maximum temperature of the warmest month for M. monachus, and the mean diurnal range and annual precipitation for S. decaocto, with a cumulative contribution rate higher than 45%.

Since the first record of M. monachus, our models predicted an increase in the current habitat suitability corresponding to 71% (5642 km²) of the MCMA territorial extension (Figure 2A), and since the first record of S. decaocto, our models predicted an increase in the current habitat suitability corresponding to 69% (5489 km²) of the MCMA (Figure 2H). Currently, the highest suitable areas for both species are mainly distributed in the north, central, east, and northeastern MCMA. These high-suitability areas are buffered by medium-suitability areas, including other areas to the west, northwest, and southeast MCMA. Unsuitable areas for both species are in the southern (e.g., Cumbres del Ajusco, Chichinautzin Biological Corridor, Izta-Popo) and western (e.g., Sierra de las Cruces) mountains surrounding the MCMA (Figure 2A,H).

3.2. Distribution Patterns Under Climate Change Scenarios

The IPSL-CM6A-LR climate change scenario projections (for the year 2050) showed slight increases and decreases in the suitable habitats for the two species, principally in areas to the north, central, and eastern MCMA (

Table 2. Area (km²) variation in the suitable habitat of Myiopsitta monachus and Streptopelia decaocto under the current and future climate scenarios.

Species	IPSL-CM6A-LR				GFDL-ESM4
	Current	SSP 126	SSP 370	SSP 585	SSP 126	SSP 370	SSP 585
Myiopsitta monachus	5642	5809	5834	5486	4273	5190	5273
	(71%)	(73%)	(73%)	(69%)	(54%)	(65%)	(66%)
Streptopelia decaocto	5489	5576	5480	5518	5134	5265	5506
	(69%)	(70%)	(69%)	(69%)	(64%)	(66%)	(69%)

, Figure 2B–G,I–N). However, under the GFDL-ESM4 climate projections, the highly suitable habitat for both species is significantly reduced by up to ~17% (Table 2, Figure 2E–G,L–N). This pattern is shown in the SSP 126 and SSP 585 scenarios (year 2050), with a slight decrease in the north and central (highest suitability), and central and western (medium suitability) peripheral areas of the MCMA for both species (Table 2, Figure 2B–G,I–N). For both species, the loss of suitable areas was slight (~24%), but steady (Figure 2B–G,I–N). The area of expansion of suitable areas for both species was less than ~800 km² (Table 2). According to our three climate change scenarios, the area of expansion for M. monachus was 167 km² for the 2041–2060 period (SSP 126) and 192 km² for the 2041–2060 period (SSP 370), while the areas of contraction ranged from 156 km² (SSP 585, IPSL-CM6A-LR) to 1369 km² (SSP 126, GFDL-ESM4) (

Table 3. Predicted changes in the climatically suitable habitat area (km²) of Myiopsitta monachus and Streptopelia decaocto under climate change scenarios.

Species		IPSL-CM6A-LR				GFDL-ESM4
		Current	SSP 126	SSP 370	SSP 585	SSP 126	SSP 370	SSP 585
Myiopsitta monachus	Stable	5642 (100%)	5642 (100%)	5642 (100%)	5486 (97%)	4273 (76%)	5190 (92%)	5273 (93%)
	Gain		167 (+3%)	192 (+3%)	-	-	-	-
	Loss		-	-	156 (−3%)	1369 (−24%)	452 (−8%)	369 (−6%)
Streptopelia decaocto	Stable	5489 (100%)	5482 (100%)	5455 (99%)	5455 (99%)	5134 (93%)	5265 (96%)	4768 (88%)
	Gain		94 (+2%)	15 (+0.3%)	63 (+1%)	-	-	738 (+13%)
	Loss		7 (−0.1%)	19 (−0.3%)	34 (−0.6%)	355 (−6%)	224 (−4%)	721 (−13%)

). In contrast, the area of expansion for S. decaocto ranged from 15 km² (SSP 370, IPSL-CM6A-LR) to 738 km² (SSP 585, GFDL-ESM4), while the areas of contraction ranged from 7 km² (SSP 126, IPSL-CM6A-LR) to 721 km² (SSP 585, GFDL-ESM4) (Table 3). Areas with stable suitability for both species were mainly located close to core zones (areas with a suitability ~ 0.75), or along discontinuous zones of suitability.

The MESS models showed different levels of dissimilarity when outside the range of the reference points (Figure 3A–N). For M. monachus, the models trained with occurrence records and projected to the MCMA showed slight climate dissimilarity in the north and west compared with those used to construct the model in the native range (Figure 3A). Under the future climate scenarios, the distribution of the climatic dissimilar area (<0, red area) in the whole potential distribution area is small (Figure 3A–G). Models for S. decaocto indicate slight climate dissimilarity in the north and central MCMA (SSP 126 in GFDL-ESM4), but this increases (SSP 585) in the western part (Figure 3H–N). Compared with the potential distribution area under the IPSL-CM6A-LR scenario in the same period in the future (Figure 3B–D,I–K), it is very noticeable in the western part, but it decreases slightly (Figure 3B–D,I–K).

3.3. Risk of Expansion into Agricultural Areas and PAs

Overall, 99.5% (3740 km²) of the current suitable habitat of M. monachus and 98% (3689 km²) of S. decaocto is within the current distribution of agricultural areas in the MCMA (Figure 4A–N). The area of suitable habitat remains stable at more than 85% for both species (Figure 4A–N). The potential suitable habitat of both species was located principally in the northern and eastern periphery of the MCMA. We identified that 80% (3711 km²) of the current suitable habitat of M. monachus is within the PAs (n = 54; 4629 km²) of the MCMA for the current period (Figure 5A–G). The area of suitable habitat remains at more than 80% for the 2041–2060 period with the IPSL-CM6A-LR climate projection (Figure 5B–D, Table S2). In contrast, the suitable area is significantly reduced to 22% in SSP 126, 46% in SSP 370, and 74% in SSP 585 under the GFDL-ESM4 climate projection scenarios (Figure 5E–G, Table S2). The potentially suitable habitats of M. monachus were located principally in PAs such as Sierra Hermosa, Zumpango lagoon, Cerro Gordo, and Sierra Pallachique (Table S2). For S. decaocto, 83% (3854 km²) of the current suitable habitat is within the PAs (n = 54; 4629 km²) for the current period (Figure 5H–N).

The area of suitable habitat remained between 64 to 84% under all scenarios (Figure 5H–N, Table S2). The potentially suitable habitats of S. decaocto were located principally in PAs such as Xochimilco, Cerro Gordo, the Cerro Ayaqueme-Huehuel volcano, Zumpango lagoon, Sierra Platachique, San Miguel Topilejo, Chichinautzin Biological Corridor, and the Tetzcotzinco system (Table S2).

A piecewise linear regression of the spread patterns in the MCMA for these species showed the presence of three relatively distinct invasion phases (Figure 6A,B). The distance from the first recorded site of invasion abruptly rose during the first year’s expansion phase from the onset (1999–2011, V = 1.45 ± 0.98 km/year for M. monachus; and 2013–2016, V = 2.67 ± 1.20 km/year for S. decaocto), with a maximum distance of 55 km in the expansion phase. The spread rates increased slightly in the second stage of the expansion phase (from 2012 to 2016 for M. monachus with V = 2.16 ± 1.01 km/year and 2017–2019 for S. decaocto with V = 3.48 ± 1.85 km/year) (Figure 6A,B), followed by a period of slow growth, with a slope not significantly different from zero (2017–2023 for M. monachus with V = 1.9 ± 0.85 km/year and 2020–2023 for S. decaocto with V = 2.28 ± 1.32 km/year), possibly representing a plateau (or saturation) phase (Figure 6A,B).

4. Discussion

Several studies over recent years have documented how the climate and land use change, and human transportation may progressively remove physiological constraints for the establishment and spread of some invasive species [103,104,105], facilitating their expansion into formerly limiting regions [35,106,107,108,109]. These processes may be accelerated via either anthropic activities or natural phenomena [10,110,111]. The model predictions for M. monachus and S. decaocto suggest that their current distributions in the invaded ranges may be largely explained either by the rapid incorporation of these species into the local culture due to their use as pets [10,14], which has probably promoted the acclimatization of these invasive species beyond their original climatic thresholds [112], or by favorable local environmental conditions [73], as is supported by previous studies demonstrating a close association of these species with human-modified habitats in North America and Europe [10,28,113,114].

Our prediction models showed that the distributions of both invasive species remain stable with slight increases (~2%), but with decreases up to 17% according to the different scenarios for the two species. The predicted invasion risk in both invasive species exhibited approximately similar trends primarily in the north, west, and central parts of the MCMA in the current and future scenarios. Our models for M. monachus showed a range increase to the periphery of the MCMA, where suitable conditions are found in greenspaces with non-native vegetation (e.g., Eucalyptus spp., Phoenix canariensis, Washingtonia robusta), and toward rural communities dedicated to grain agriculture in the case of S. decaocto, suggesting these habitats as favorable for the establishment of these species. These results also support previous findings suggesting that these species may be limited to human-altered habitats in newly invaded areas [107,108,109,115]. However, some studies have suggested that S. decaocto tends to avoid areas of intensive agriculture and conserved forest [28,116,117].

In Europe, the distribution of M. monachus and S. decaocto are generally restricted to relatively warmer regions, although both species have also been found in temperate areas [28,113,118,119]. Romagosa and McEneaney [8] and Ingeloff et al. [73] mentioned that these species have successfully colonized tropical, subtropical, temperate, and arid areas in North America and the Caribbean region, which suggests that these species may track environmental conditions similar to those in their original distribution area in southern Asia [21,22]. These findings suggest that in Mediterranean Europe, individuals may come from local breeding populations, reflecting a process of a posteriori population establishment (i.e., when a small group of individuals separate from the main population to form a new colony), as shown in other invasive species [10,113,114,120,121].

Hiley et al. [122] cautioned that while PAs may not facilitate the colonization of invasive species at their initial stage, they are more susceptible to invasion and an exponential population increase. Although our results showed that both species may be in a saturation phase (or plateau), these invasive bird species represent a risk to conserved environments around the MCMA due to their rapid spread, which may be enhanced through the growth of the urban area and reforestation within the PAs with exotic trees that facilitate their establishment. Local studies may support this process in PAs in different regions of Mexico [109,115,123,124].

Some invasive alien species may remain in a quiescence state, referred to as a lag phase [125,126], after their introductions, occupying a small area for several years before rapidly expanding. However, the populations of M. monachus and S. decaocto in invaded areas have grown exponentially over short time periods [10,127,128,129]. For example, S. decaocto has experienced three phases during its colonization and establishment in North America: (1) a population exponential increase, (2) a leveling at or above the local environment carrying capacity, and (3) an a posteriori decrease [130]. An initial lag phase, associated with the first population establishment of these species in the MCMA but with limited spatial expansion, was not clearly visible from our data; it may have taken place before the first species record, or may have been very short, or it is unclearly defined because of multiple intentional or unintentional releases, as these birds are maintained as pets. However, it would be necessary to consider whether other factors, such as a low abundance, a death rate exceeding the birth rate, and competition between native species or with other already established invasive species, may have been overcome. This may suggest that the populations of these species routinely overshoot the carrying capacity of local habitats. Our results suggest that the relatively recent records in central Mexico present this pattern, suggesting that these species have been expanding their range distribution in the second stage of the expansion phase, but still have not yet reached the a posteriori decrease stage.

From this perspective, the distribution expansion of M. monachus and S. decaocto in central Mexico may be largely attributed to their propensity to move and occupy suitable conditions in contiguous areas favoring their establishment [11,14,113,116]. However, M. monachus and S. decaocto can presumably “jump” larger distances if suitable conditions are not available at close ranges, thus starting the colonization process, which may be followed by a process of establishment, or it may be that they have been released in nearby areas and thus simulate long-distance dispersal [73,113,118].

Our models support a slight increase in the distribution of M. monachus and S. decaocto in the invaded regions, suggesting a lack of acclimatization to changing environments, and thus decreasing the significance of pets being released. In this sense, our results are largely consistent with those of previous studies [131,132,133], in which a suitable habitat range will slightly increase as global warming intensity increases under low-level SSPs, but in which excessive temperature increases might restrict or decrease the habitat suitability [134,135].

Future projections of our models for both species showed that the climatic suitability remains relatively stable but presents considerable decreases up to 24% for M. monachus and up to 17% for S. decaocto [see Table 3]. This pattern is congruent with invasive phases, where decreases noted in areas of climatic suitability are possibly related to a plateau or saturation phase, after an increase that seems related to the second stage of the expansion phase. Then, the expansion phase flows into a saturation phase with extremely low spread rates and a lower (but still increasing) occupancy range because of biological constraints and/or the filling of suitable niche environments.

5. Conclusions

Invasive alien species may have major negative impacts on native communities and promote the homogenization of flora and fauna [36,136,137,138]. However, although M. monachus and S. decaocto are both listed as a highly invasive species [139,140], they are still being traded as pets in central Mexico, as well as in other regions of the country. However, although the number of established populations is apparently still small, and the potential damage is still unknown [141], timely and efficient monitoring is necessary to apply rapid and effective actions to contain or eradicate these invasive alien species. Otherwise, in the short term, these species may potentially colonize sites with limited anthropic influence [142,143], such as PAs [103,144]. Given that population dynamics and the invasive potential of these species in Mexico are still largely unknown, it is essential that government agencies finance and promote further scientific research to establish the effect of its presence in Mexico. Therefore, an integrated invasive species management technique should focus on eradication or population reduction (e.g., using toxicants, shooting, and trapping) to keep populations at levels where nonlethal tools can be utilized to reduce damage. However, the efficacy of an eradication campaign depends on biological, environmental, and economic factors, along with the issuance of a social license for lethal removal [145]. In addition, environmental education programs and dissemination via different media are required, regarding the danger posed by the introduction of invasive alien species and the impacts these invasive species may have on the biota and native ecosystems.

Although biological invasions are now better understood from an ecological perspective, a clearly less studied issue is the human dimension. Human activities could artificially expand the geographical range [146,147] of M. monachus and S. decaocto. It is possible that range expansion has been promoted due to a population increase in towns surrounded by agricultural areas, which benefited its presence, likely due to stored food sources [10,28,116,117,127].

Climate niche models are highly important tools in the study of prevention and control of biological invasions [148,149,150], but their application requires accurate predictions on the potential spatial extent of these invasions, as this facilitates concentrating economic and research efforts aimed at preventing the arrival of invasive species. In addition, potential distribution and climate niche models of invasive species also provide new opportunities to analyze whether their expansion depends on environmental conditions or human activities.