Factors affecting presence and relative abundance of the Endangered volcano rabbit Romerolagus diazi, a habitat specialist
F ELIPE O SUNA,ROGER G UEVARA,ENRIQUE M ARTÍNEZ-MEYER R AÚL A LCALÁ and A LEJANDRO E SPINOSA DE LOS M ONTEROS
Abstract Habitat specialists are particularly vulnerable to factors limiting a species’ distribution. Habitat generalists extinction when habitat conditions are altered. Information can tolerate a broad spectrum of environmental conditions, on the habitat use of such species is thus important because whereas specialists are restricted to a limited set of environ- it provides insight into factors that influence distribution and mental conditions, which restricts their areas of distribution abundance, which is crucial for conservation. Here, we aimed and makes them more vulnerable to disturbance and loss of to identify factors that influence the patterns of presence specific habitat resources (Crown & van Riper, ). Where and abundance of the Endangered volcano rabbit Romerolagus a species lies along this generalist–specialist continuum diazi, a rare leporid with a patchy distribution. Through ex- affects its population dynamics (Kolasa & Li, ). In haustive sampling of its range in the Sierra Chichinautzin the current context of global environmental change (e.g. and Sierra Nevada volcanic fields, Mexico, and using general- through climate change, habitat loss and pollution) it is a ized linear models, we found that the probability of patch priority to document habitat use by specialist species, and to occupancy was higher where bunchgrass cover exceeded consider natural and anthropogenic factors that influence %, rock cover exceeded %, no cattle grazing was observed their distribution. Such studies facilitate the identification and human settlements were at least km away. Patches with not only of habitat units within the landscape that could greater relative abundance were those with similar character- potentially support viable populations, but also of factors istics, but located at elevations . , m, and with rock that limit the species’ distribution. In particular, this informa- cover , %. Cattle grazing was identified as a major threat tion is fundamental for mitigation measures that encourage to local populations of the volcano rabbit, particularly in habitat occupation, which is a priority for the conservation the Sierra Chichinautzin. Because of the significance of of threatened species (Edgel et al., ). bunchgrasses for this species, the protection of the mountain The volcano rabbit Romerolagus diazi (also known as grasslands is required in both volcanic fields. zacatuche or teporingo) is a mountain habitat specialist en- demic to the Sierra Chichinautzin and Sierra Nevada volcan- Keywords Anthropogenic disturbance, bunchgrass cover, ic fields in the Trans-Mexican Volcanic Belt. It is associated cattle grazing, conservation, Leporidae, Romerolagus diazi, with patches containing bunchgrasses (e.g. Muhlenbergia volcano rabbit spp. and Festuca spp.) that have a fragmented distribution Supplementary material for this article is available at within forests at elevations of ,–, m (Velázquez doi.org/./S et al., ). There is increasing anthropogenic pressure on volcano rabbit populations. Cattle grazing, logging and land conversion for agriculture lead to habitat modification, fragmentation and loss (Galicia & García-Romero, ; Introduction Arregoitia et al., ). As a result, the volcano rabbit is ca- wo major challenges in studying wildlife ecology are the tegorized as Endangered on the IUCN Red List (Velázquez characterization of habitat use and the identification of & Guerrero, ), and considered in danger of extinction T by Mexican laws (SEMARNAT, ). High abundance of this lagomorph has been recorded in FELIPE OSUNA ( orcid.org/0000-0003-2148-6583), ROGER GUEVARA and sites covered by grass species with a bunch type growth habit ALEJANDRO ESPINOSA DE LOS MONTEROS (Corresponding author, orcid.org/ – 0000-0002-7742-5977) Departamento de Biología Evolutiva, Instituto de (Festuca tolucensis and Festuca tolucensis Trisetum altiju- Ecología AC, Carretera Antigua a Coatepec 351, El Haya, Xalapa 91070, gum), followed by bunchgrasses associated with pine forests Veracruz, Mexico. E-mail [email protected] (Muhlenbergia quadridentata and F. tolucensis associated ENRIQUE MARTÍNEZ-MEYER Instituto de Biología, Universidad Nacional with Pinus hartwegii; Velázquez & Heil, ). However, Autónoma de México, Ciudad Universitaria, Mexico City, Mexico the rabbit’s preference for biological and structural habi- RAÚL ALCALÁ Departamento de Ecología Evolutiva, Centro de Investigación en tat features varies locally. On the Pelado volcano in the Biodiversidad y Conservación, Universidad Autónoma del Estado de Morelos, Morelos, Mexico Sierra Chichinautzin, rabbit populations occurred mainly – Received June . Revision requested September . in mixed forests (Pinus spp. Alnus spp.; Fa et al., ). On Accepted April . some volcanoes in the southern Sierra Chichinautzin, at
This is an Open Access article, distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike licence (http://creativecommons.org/licenses/by-nc-sa/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the same Creative Commons licence is included and the original work is properly cited. DownloadedThe from written https://www.cambridge.org/core permission of Cambridge University. IP address: Press must 170.106.202.126 be obtained for, on commercial 28 Sep 2021 re-use. at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/termsOryx, Page 1 of 10 © The Author(s),. https://doi.org/10.1017/S0030605320000368 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 2 F. Osuna et al.
elevations of ,–, m, the species’ abundance was in- activities (Galicia & García-Romero, ). The forest is the fluenced by the patch size and cover extent of bunchgrasses, main source of income and energy (i.e. fuel for cooking and and the presence of shrubs (Rizo-Aguilar et al., , ; heating) for local communities (Trejo & Dirzo, ). There Uriostegui-Velarde et al., ). On the Iztaccíhuatl volcano are four main land-use types in the region: cropland, in the Sierra Nevada (,–, m), forest cover and cat- pastures for livestock grazing, forests (used for timber and tle grazing correlated negatively with habitat occurrence for firewood extraction) and urban areas. Other activities this species, whereas population abundance increased with include hunting and the extraction of fertile soil that is distance from the surveillance facilities of the Izta-Popo sold for use in gardening (Velázquez et al., ; Galicia & National Park (Hunter & Cresswell, ). García-Romero, ). All previous studies were carried out in specific locations Despite their geographical proximity, there are profound at restricted elevation ranges. It is unknown whether the dif- differences between the two mountain ranges. Habitat ferences observed between sites are a consequence of specif- patches in the Sierra Chichinautzin are small and isolated ic local conditions or constitute a general pattern of habitat by a forest matrix with agricultural fields, with high levels use by the volcano rabbit across its range. Therefore, we con- of human disturbance (forest exploitation and cattle grazing; sidered the entire known range of this species, including the Table ). Environmental conditions in the Sierra Chichinautzin two major volcanic ranges, its full elevational range and all are warmer and drier than in the Sierra Nevada and thus less suitable habitats in these landscapes. We aimed to answer favourable for the growth of bunchgrasses. In addition, the following questions: () which factors (biotic, abiotic urban pressure is higher because the Sierra Chichinautzin and anthropogenic) influence patch occupancy and relative is near Mexico City and Cuernavaca. In the Sierra Nevada abundance of the volcano rabbit, and () to what extent is patches of grassland are larger, more connected and less cattle grazing a threat to its populations? Given the known affected by human activities, presenting more favourable association of volcano rabbits with bunchgrasses, we pre- conditions for the volcano rabbit. dicted: () a positive correlation between both patch occu- pancy and relative abundance with per cent bunchgrass cover, and ( ) reduced patch occupancy and relative abun- Methods dance in areas grazed by cattle. Survey design
Study area Based on historical records of the volcano rabbit (Velázquez et al., ; Domínguez, ; Osuna et al., ) we delim- This work was carried out in the Sierra Chichinautzin and ited the possible range of the species in both volcanic fields Sierra Nevada volcanic fields. These mountains are located as areas at ,–, m altitude. During and we in the eastern Trans-Mexican Volcanic Belt, adjacent to surveyed a total of km in search for patches with dense Mexico City, one of the world’s largest urban centres. bunchgrass cover that could be suitable habitat for the Together, these volcanic fields encompass the global range volcano rabbit (Velazquez & Heil, ). In areas with frag- of the species (Hunter & Cresswell, ). The Sierra Chichi- mented habitats, we considered bunchgrass patches that nautzin comprises more than volcanoes, whereas the were at least m apart as independent sites. We found Sierra Nevada consists of only five large volcanoes, including independent sites of bunchgrass patches adequate for Popocatépetl (, m) and Iztaccíhuatl (, m), the second the presence of volcano rabbits. We marked a × m and third highest in Mexico (Márquez et al., ). Vegetation quadrant in each patch (Mueller-Dombois & Ellenberg, cover is temperate forests (e.g. Quercus spp., Pinus spp. and ), to determine the presence of volcano rabbits and Abies spp.) and subalpine and alpine grasslands (e.g. Muhlen- gather data about environmental variables considered to bergia spp., Festuca spp. and Calamagrostis spp.; Rzedowski, be associated with habitat quality. We placed the sample ). We observed seven species of bunchgrasses that can quadrant near the centre of each patch, avoiding roads or be used by the volcano rabbit: Muhlenbergia macroura, M. areas that showed clear evidence of human alteration. We quadridentata, F. tolucensis, Festuca rosaceae, Festuca amplis- placed metal posts at the quadrant corners and marked the sima, Calamagrostis tolucensis and Stipa ichu. The dominant perimeter with a rope tied to the corner posts. Using cotton species were M. macroura and F. tolucensis,whichwerepre- thread tied to the perimeter ropes at m intervals, we divided sent in . % of the patches occupied by the volcano rabbit. each quadrant into m grid cells, to take fine-scale measure- TheSierraChichinautzinandSierraNevadaharbouragreat ments of environmental variables. diversity of vertebrates, including some micro-endemic high The volcano rabbit is secretive and shy, hiding in bur- mountain specialists such as the volcano axolotl Ambystoma rows at any perceived danger, which means direct obser- leorae (Monroy et al., ). There are numerous human set- vation is rarely possible. We therefore used the indirect tlements in the area, and associated agricultural and forestry method of excreta presence to infer patch occupancy. This
Oryx, Page 2 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368 https://www.cambridge.org/core/terms Downloaded from Oryx ae3o 10 of 3 Page , https://www.cambridge.org/core © h uhrs,22.Pbihdb abig nvriyPeso eafo an lr nentoa.doi:10.1017/S0030605320000368 International. Flora & Fauna of behalf on Press University Cambridge by Published 2021. Author(s), The . https://doi.org/10.1017/S0030605320000368 . IPaddress: TABLE 1 Variables influencing the presence and relative abundance of the volcano rabbit Romerolagus diazi throughout its range in the Sierra Chichinautzin and the Sierra Nevada, Mexico.
Unoccupied patches Occupied patches 170.106.202.126 Sierra Chichinautzin Sierra Nevada Total Sierra Chichinautzin Sierra Nevada Total Variable No./Mean ± SE No./Mean ± SE No./Mean ± SE No./Mean ± SE No./Mean ± SE No./Mean ± SE Analysed patches 53 41 94 44 27 71 No. of latrines 0 0 0 54.8 ± 27.6 65.2 ± 37.6 58.8 ± 31.9 , on Altitude1,2 (m) 3,204 ± 194 3,626 ± 314 3,388 ± 328 3,142 ± 151 3,705 ± 302 3,356 ± 351 28 Sep2021 at19:39:59 Bunchgrass cover1,2 (%) 78.4 ± 16.4 78.6 ± 13.9 78.5 ± 15.3 82.8 ± 9.0 82.9 ± 11.0 82.9 ± 9.7 Cattle grazing1,2 22 21 43 6 8 14 Forest cover1,2 (%) 22.7 ± 23.6 29.9 ± 27.0 25.8 ± 25.3 25.2 ± 18.3 13.7 ± 17.1 20.8 ± 18.6 Herbaceous cover (%) 7.6 ± 11.1 7.6 ± 8.6 7.6 ± 10.4 4.4 ± 4.6 6.5 ± 8.3 5.2 ± 6.2 Human impact index 0.8 ± 0.2 0.5 ± 0.2 0.6 ± 0.2 0.7 ± 0.2 0.6 ± 0.1 0.7 ± 0.2 Distance to nearest settlement1 (km) 4.8 ± 2.0 7.9 ± 2.8 6.1 ± 2.2 5.1 ± 2.3 8.1 ± 2.9 6.2 ± 2.9 , subjectto theCambridgeCore termsofuse,available at Rock cover1,2 (%) 2.3 ± 5.4 1.6 ± 5.7 2.0 ± 5.5 5.8 ± 7.4 1.7 ± 5.1 4.2 ± 6.9 Shrubs cover1,2 (%) 3.7 ± 6.2 2.1 ± 5.2 3.0 ± 5.8 3.7 ± 4.9 4.1 ± 5.9 3.8 ± 5.3 Terrain slope 9.9 ± 7.9 13.0 ± 8.0 11.2 ± 8.0 8.7 ± 6.5 12.5 ± 6.6 10.1 ± 6.8 Sylvilagus cunicularius presence1,2 23 10 33 10 6 16 Sylvilagus floridanus presence 13 2 15 8 6 14 No. of mature trees3 1.2 ± 1.9 2.5 ± 3.5 1.8 ± 2.8 1.2 ± 1.6 0.8 ± 1.2 1.1 ± 1.5
rabbit volcano The Variable used in presence model. Variable used in relative abundance model. Diameter at breast height . cm. oeoau diazi Romerolagus 3 4 F. Osuna et al.
is a widely used method for the study of this and other lago- the human impact index on Mexican biodiversity from morphs (Krebs et al., ; Hunter & Creswell, ). We did MEXBIO (CONABIO, ), which is based on the theo- not consider individual droppings nor dispersed excreta as retical framework of the Global Biodiversity Model and evidence for rabbit presence, but searched the entire ground establishes simple cause–effect relationships between dif- surface of the quadrant for latrines. Latrines are defined as ferent environmental pressures and threats. We adapted the areas with more than fresh faecal pellets grouped together model for Mexico at a resolution of km , including factors (Santilli et al., ), are easy to detect and indicate recent such as land-use pressure, infrastructure and habitat frag- occupation by volcano rabbits. Patches with burrows but mentation. The distance to human settlements affects local without latrines could be abandoned sites. We recorded a conditions such as habitat quality, whereas the index of site as inhabited by volcano rabbits only when both latrines human impact mainly influences spatial processes such as and burrows were present. One latrine is not necessarily species dispersal amongst patches. equivalent to one individual, but a positive relationship be- tween the number of latrines and the number of individuals has been suggested (Hunter & Creswell, ; Santilli et al., Data analysis ). Therefore, recording the number of latrines is an in- We performed statistical analyses in R .. (R Development direct way of inferring relative abundance. Core Team, ). We created two matrices: () presence analysis, to detect variables that influenced patch occu- pancy, and () relative abundance analysis, to detect vari- Variables related to habitat use ables that influenced this parameter. For each matrix we constructed generalized linear models. To assess presence, We recorded a total of variables related to habitat for we built a binomial regression model (presence/absence) analysis (Table ), based on the known natural history of with the data from all patches, using the logit link function the species (Velázquez et al., ) and previous studies on in R. For relative abundance, we constructed a model based its habitat use (Fa et al., ; Hunter & Creswell, ; only on data from occupied patches, using the gamma Rizo-Aguilar et al., ). To analyse vegetation cover we distribution and reciprocal link functions in R. Given the measured the per cent cover of bunchgrasses higher than models’ complexity ( variables and their interactions), cm, herbaceous plants and shrubs, and counted the num- we followed a series of steps to find the simplest possible ber of adult trees with a diameter at breast height of . cm. model: () We explored each variable using a univariate We quantified forest cover using a spherical densiometer, model, to infer its individual contribution. () For each estimating the value for each site as the mean of the four matrix we reduced the number of variables depending measurements corresponding to the different cardinal points on the individual contribution of each variable (Table ). from the centre of the quadrant. In some instances, different () We evaluated models with all paired interactions be- vegetation types (e.g. grass, shrubs, herbaceous plants) over- tween their variables. () We performed a stepwise bidi- lapped, so the sum of their cover percentages was . % rectional simplification procedure employing the Akaike of the quadrant area. We recorded the presence of other information criterion (AIC) value as a criterion for model rabbits (the Mexican cottontail Sylvilagus cunicularius and selection, using the stepAIC function in the MASS package eastern cottontail Sylvilagus floridanus)byidentifyingtheir for R.() We refined the simplification by comparing mod- excreta, which are easily distinguished from those of the els using the difference in variance until we identified the volcano rabbit (Velázquez et al., ). simplest model. We extracted abiotic variables, including per cent cover of rocks, elevation and slope, from the official Mexican Digital Elevation Model (Mexican Elevation Continuum, Results CEM .,at m; INEGI, ) using QGIS . (QGIS Development Team, ). The identity of the volcanic field We detected volcano rabbits in of the patches visited (i.e. Sierra Chichinautzin or Sierra Nevada) was included in (Fig. ). The model that best explained their presence re- theanalysisasanabioticfactor. tained nine variables and interactions, accounting for We determined the presence of cattle grazing by livestock % of the total variance. Bunchgrass cover, rock cover, sightings, the presence of excreta, and evidence of grazing at distance to the nearest human settlement and cattle graz- or close to the bunchgrass patch. We calculated the distance ing were the variables with the most significant effects from the centre of each sample patch to the nearest human (Table ). The probability of patch occupancy increased settlement using Google Earth Pro (US Department of State with higher per cent cover of bunchgrasses and rocks Geographer, ). This programme accounts for relief and (Fig. a,b), but decreased in the presence of cattle grazing landscape features such as creeks or hills when measuring and in closer proximity to human settlements (Fig. c,d). distances on the Earth’s surface. We included in our analysis However, the effect of the distance to human settlements was
Oryx, Page 4 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368 The volcano rabbit Romerolagus diazi 5
FIG. 1 Current distributional range of the volcano rabbit Romerolagus diazi in the Sierra Chichinautzin and the Sierra Nevada in central Mexico.
different between the two volcanic fields (Supplementary of volcano rabbits was significantly higher in patches with Fig. a). In the Sierra Chichinautzin, the probability of . % of bunchgrass cover (Fig. a), and in those located patch occupancy was , % when human settlements were at altitudes . , m(Fig. b). Relative abundance had a , km away (P , .), but increased rapidly as distance negative relationship with rock cover (Fig. c) and cattle increased (P , .). In the Sierra Nevada, we observed grazing (Fig. d). Regarding interactions between variables, no significant effect for this factor (P = .). we detected that volcano rabbits were most abundant in In the Sierra Chichinautzin, the presence of cattle graz- patches with . % bunchgrass cover, , % shrub cover ing reduced the probability of patch occupancy to %, and . % forest cover (Supplementary Fig. a,b). We whereas without cattle grazing it was up to %. However, found an increased relative abundance of volcano rabbits cattle grazing did not change the probability of patch occu- as forest cover increased, if rock cover was . % (Supple- pancy in the Sierra Nevada, where it remained at c. % mentary Fig. c). The effect of bunchgrass cover was great- (Supplementary Fig. f). Nevertheless, the variance of the er in the Sierra Nevada than in the Sierra Chichinautzin probability for patch occupancy increased where cattle graz- (Supplementary Fig. f). Above , m, shrub cover had ing occurred. The interaction between cattle grazing and a positive (Supplementary Fig. d) and rock cover a negative distance to human settlements had a negative effect on effect on relative abundance (Supplementary Fig. e). In the patch occupancy. Volcano rabbits were nearly absent from Sierra Chichinautzin the presence of the Mexican cottontail cattle-grazed patches located less than km from human negatively affected the relative abundance of the volcano settlements, but the probability of patch occupancy in- rabbit, whereas in the Sierra Nevada the effect of such creased rapidly with increasing distance from human set- sympatry was positive (Supplementary Fig. g). tlements (Supplementary Fig. b). In the absence of cattle grazing, the distance to human settlements did not affect the probability of patch occupancy. The interaction between Discussion presence of the Mexican cottontail and per cent shrub cover was significant. When another lagomorph was pre- The probability of recording volcano rabbits and their rela- sent, the probability of recording the volcano rabbit increased tive abundance in occupied areas was associated with dense as shrub cover increased (Supplementary Fig. c). An increase cover of native bunchgrasses and rocky areas at high eleva- in rock cover and forest cover (Supplementary Fig. e,f) tions on the volcanoes surrounding Mexico City. The prox- correlated positively with patch occupancy at elevations imity of these mountains to one of the most populated , , m, where these elements are common. areas in the world leads to severe habitat modifications The model that best explained relative abundance re- (e.g. natural resource extraction, introduction of livestock, tained eight variables and seven interactions, explaining accidental and provoked fires) that negatively affect the % of the total variance (Table ). The relative abundance volcano rabbit and many other wildlife species.
Oryx, Page 5 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368 6 F. Osuna et al.
TABLE 2 Summary of effects in the generalized linear model fitted to presence (probabilistic distribution binomial, logit link function) and relative abundance (probabilistic distribution gamma, inverse link function) of the volcano rabbit. Variables marked with * are significant for both parameters.
Presence Abundance Variable χ2 P χ2 P Altitude 0.024 0.8760 7.119 0.0081 Bunchgrass cover* 5.357 0.0206 5.213 0.0224 Cattle grazing* 16.378 , 0.0001 7.852 0.0056 Forest cover 0.820 0.3652 0.011 0.9127 Distance to nearest settlement 5.239 0.0221 Shrubs cover 2.996 0.0835 3.711 0.0524 Sierra 0.026 0.8714 1.954 0.1560 Sylvilagus cunicularius 1.003 0.3166 0.629 0.4178 Rock cover* 11.549 0.0007 4.742 0.0291 Altitude × Cattle grazing 0.638 0.4246 Altitude × Forest cover 11.687 0.0006 Altitude × Rock cover* 18.012 , 0.0001 9.515 0.0025 Altitude × Shrubs cover 37.376 , 0.0001 Altitude × Sierra 1.710 0.1909 Altitude × Sylvilagus cunicularius 1.288 0.2565 Bunch grass cover × Forest cover 8.523 0.0040 Bunch grass cover × Rock cover 1.964 0.1611 Bunch grass cover × Shrubs cover 9.179 0.0029 Bunch grass cover × Sierra 10.291 0.0017 Cattle grazing × Distance to settlements 11.599 0.0007 Cattle grazing × Sierra 13.836 0.0002 Cattle grazing × Sylvilagus cunicularius 1.340 0.2471 Forest cover × Rock cover 9.296 0.0027 Distance to settlements × Sierra 8.118 0.0044 Rock cover × Shrubs cover 1.932 0.1645 Shrubs cover × Sylvilagus cunicularius 9.905 0.0016 Sierra × Sylvilagus cunicularius 15.574 0.0002
Biotic factors Volcano and pygmy rabbits are the two smallest leporid species, both with a mean weight , g (Green & Flinder, Bunchgrasses are a primary resource that defines habitat ; Cervantes et al., ). suitability for the volcano rabbit throughout its range; this Sympatry between volcano rabbits and other Mexican is consistent with previous local reports (Velázquez et al., rabbits seems to be mediated by shrub cover. The presence ; Hunter & Creswell, ) and can be explained by of shrubs may indicate better habitat quality, and therefore the species’ ecology. Bunchgrasses represent . % of the allow the co-occurrence of species sharing similar niches. volcano rabbit’s diet (Cervantes et al., ; Martínez- Another possibility is that competition for bunchgrasses García et al., ). As the rabbits move through the habitat, may force volcano rabbits into areas with denser shrub the grasses are bent, forming tunnels and passages that pro- cover. To assess the likelihood of these scenarios, it would vide cover from predators (Cervantes, ). These runways be necessary to carry out studies that include experimental are marked with latrines to define territories. The rabbits manipulation (e.g. shrub removal, shrub cover modifica- also use bunchgrasses as nest-building material, and the tion). At elevations . , m the relative abundance of base of well-developed bunches as nests (Velázquez et al., the volcano rabbit was higher in patches with denser ). The model constructed for relative abundance de- shrub cover. It has been reported that during times of tected that the interactions of bunchgrass cover with harsh environmental conditions (cold and dry winter shrub cover, forest cover and identity of the volcanic field months when grasses are scarce), volcano rabbits increase were significant, but in all cases the relative abundance their consumption of shrubs (shrubs comprise .%of was greater in those patches that had bunchgrass cover their diet in the rainy season and up to .% in the dry sea- . %. This degree of habitat specificity is comparable to son; Cervantes & Martínez, ). This increased consump- that of the pygmy rabbit Brachylagus idahoensis. This spe- tion of shrubs could also occur in the more restrictive cies is distributed in the western USA and is associated environmental conditions near the upper elevation limit with plants of the genus Artemisia (Schmalz et al., ). of the species’ distribution.
Oryx, Page 6 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368 The volcano rabbit Romerolagus diazi 7
FIG. 2 Influence of the main variables on patch occupancy by volcano rabbits. The proportion of bunchgrass (a) and rock cover (b) were the essential factors that made a patch suitable for the volcano rabbit. Proximity to human settlements (c) and livestock grazing decreased volcano rabbit occupancy (d).
FIG. 3 Influence of the variables on relative abundance of the volcano rabbit. The proportion of bunchgrass cover shows a positive relationship with the relative abundance (a). Bunchgrass is a vital resource that is used as food, protection against predators, and as nesting material. As altitude increases, tree cover and abundance of other plants decrease, leaving patches composed mainly of bunchgrasses and increasing the relative abundance of rabbits (b). Although rocks are another basic resource, a high per cent of rock cover reduces the area available for bunchgrasses and thus reduces the relative abundance of volcano rabbits (c). Relative abundance is affected negatively by the presence of livestock (d).
Sympatry with the Mexican cottontail was negatively cor- in the Sierra Chichinautzin, which could intensify com- related with the relative abundance of the volcano rabbit petition between the volcano rabbit and other lagomorphs in the Sierra Chichnautzin, but positively correlated in the (Leach et al., ). This finding is consistent with high levels Sierra Nevada. Higher levels of human disturbance (human of physiological stress (inferred by the amount of cortisol impact index of . in the Sierra Chichinautzin vs . in and its derivatives measured in the excreta) reported for the the Sierra Nevada; Table ) may indicate lower habitat quality volcano rabbit in some patches in the Sierra Chichinautzin
Oryx, Page 7 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368 8 F. Osuna et al.
(Rizo-Aguilar et al., ). In the larger areas of grassland patches of suitable habitat have rocks and the occupied away from human influence in the Sierra Nevada the rabbits patches averaged .% of rock cover. The existence of larger may experience less pressure on resources such as food and and more connected patches in the Sierra Nevada compared shelter. It is also possible that sympatry with other lago- to the Sierra Chichinautzin leads to a greater probability of morphs could be positive for the volcano rabbit in these detecting the presence of volcano rabbits even when the areas because of reduced predation pressure. However, stud- rock cover in some of those patches is , %. The positive ies with an experimental design would be required to ex- effect of forest cover on volcano rabbit presence at elevations amine these interspecific interactions more thoroughly. , , m can be explained by the fact that forest can be ex- tensive at those elevations, whereas at higher elevations for- Abiotic factors est cover is naturally lower. In both sierras, large proportions of forest cover may indicate low levels of human disturb- Rock cover was determinant for the presence and relative ance, and thus an increased probability of species detection. abundance of the volcano rabbit. The probability of patch In addition, it has been reported that volcano rabbits feed on occupancy was highest where rock cover was –% but leaves of Alnus spp. (Cervantes & Martínez, ), which is the relative abundance of rabbits in those patches was common in both sierras at elevations , , m and contri- lower than in occupied patches with rock cover , %. For butes significantly to forest cover. small mammals such as the volcano rabbit, rocks may act as refuges that facilitate predator avoidance (Alves et al., ). Anthropogenic factors The species exhibits social aggregation and is the only lep- orid that vocalizes (Velázquez et al., ). Sometimes it is We found that cattle grazing and close proximity to human possible to observe individuals on rocks, keeping watch settlements significantly reduced occupancy by volcano rab- and emitting alerts to the group if a threat is detected. bits, particularly in the Sierra Chichinautzin. This suggests However, sites with a rock cover . % are characterized that this species, like other habitat specialists, is sensitive by a low relative abundance of volcano rabbits, as a greater to human disturbance. Human impact on ecosystems leads per cent of rock cover means less bunchgrass cover. to the loss of specific resources, in this case bunchgrasses. The effect of elevation on relative abundance reflects the A negative effect of anthropogenic disturbances on habitat species’ habitat preference, considering that bunchgrasses pre- specialists has been demonstrated in other studies (Pierce dominate at elevations . , m(Velázquez&Heil,; et al., ; Dendup et al., ; Zhao et al., ). We Velázquez et al., ). Typically, at elevations , , m found that the presence of livestock grazing primarily habitat patches are small and with substantial tree cover affected patch occupancy, but had little effect on relative (Abies spp., Alnus spp. and Pinus spp.), isolated by a forest abundance. It is possible that this could be the effect of matrix mixed with agricultural fields. Large habitat patches recent cattle introduction to patches that have long been (with high relative abundance of volcano rabbits) can also used by rabbits. be found at altitudes of ,–, mingrasslandswith Cattle feed mainly on bunchgrasses, and the way in pine forest, characterized by the predominance of M. macro- which they graze reduces grass cover and alters its natural ura–Pinus spp. or F. tolucensis–Pinus spp. At elevations structure. High rates of fires are also associated with grazing . , m,thereareextensiveareasofsubalpineand because of prescribed burning during the dry season to alpine grasslands (e.g. Festuca spp., Muhlenbergia spp. and stimulate pasture regrowth (Blanco et al., ). Fires are a Calamagrostis spp.), with large habitat patches in a matrix major threat to lagomorph conservation (Portales et al., that is permeable to movements of volcano rabbits. Our ). Burning pasture is a common activity throughout the findings differ from those of Velázquez & Heil (), who re- entire range of the volcano rabbit, causing numerous un- ported a negative correlation between elevation and the rela- controlled fires that devastate hundreds of hectares of grass- tive abundance of volcano rabbits in the Sierra Chichinautzin. lands per year. Other factors that reduce habitat quality in This discrepancy could be attributable to sampling design: the the vicinity of human settlements (Velázquez et al., ) highest elevations of the Sierra Chichinautzin are , , m, include agriculture, wood extraction, removal of fertile soil, so there may have been few suitable patches for the volcano hunting, road construction, pollution and presence of feral rabbit at the study site of Velázquez & Heil (). fauna. These factors need to be taken into account when The interactions between rock cover and elevation, and planning and implementing conservation actions. between rock cover and forest cover, appeared to be a result of spatial autocorrelation between these factors. At eleva- Implications for conservation tions , , m, as in the Sierra Chichinautzin, %of the patches of suitable habitat have rocks, and the mean Our findings can contribute to the design and development rock cover in occupied patches was .%. At elevations of effective conservation plans for the volcano rabbit and . , m, as in most of the Sierra Nevada, % of the its habitat. The species’ dependence on bunchgrasses makes
Oryx, Page 8 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368 The volcano rabbit Romerolagus diazi 9
the protection of the natural grasslands in the Sierra CERVANTES, F.A. & MARTÍNEZ,J.() Food habits of the rabbit Chichinautzin and the Sierra Nevada a matter of high Romerolagus diazi (Leporidae) in central Mexico. Journal of – importance. We found that the presence of livestock and Mammalogy, , . CERVANTES, F.A., LORENZO,C.&HOFFMANN, R.S. () probably other factors such as fires and agricultural expan- Romerolagus diazi. Mammalian Species, , –. sion may represent a threat to this key resource, with grazing CERVANTES, F.A., LORENZO, C., VARGAS,J.&HOLMES,T.() directly affecting bunchgrass cover and structure. Mitigating Sylvilagus cunicularius. Mammalian Species, , –. the impact of cattle grazing is a priority at both volcanic CONABIO () Modelo Espacial de Impactos a la Biodiversidad fields because of its negative effect that can lead to local ex- Mexicana, MEXBIO. Resolución km . Comisión Nacional para el Conocimiento y Uso de la Biodiversidad, Mexico City, Mexico. tirpation of volcano rabbit populations. Furthermore, the CROWN,C.&VAN RIPER,C.() Avian community responses to lack of trained personel to enforce the current regulations mechanical thinning of a pinyon-juniper woodland: specialist (SEMARNAT, ; DOF, ) limiting the exploitation sensitivity to tree reduction. Natural Areas Journal, , –. of natural resources is a serious problem. In addition to in- DENDUP, P., CHENG, E., LHAM,C.&TENZI,U.() Response of the creasing capacity for law enforcement, environmental edu- Endangered red panda Ailurus fulgens fulgens to anthropogenic disturbances, and its distribution in Phrumsengla National Park, cation programmes should be established. These measures Bhutan. Oryx, , –. could support the conservation of not only the volcano DOF (DIARIO OFICIAL DE LA FEDERACIÓN)() Ley General de rabbit, but also other wildlife species in these ecosystems. Vida Silvestre. Diario Oficial de la Federación. January , An imminent threat that has not been evaluated is af- Mexico City, Mexico. forestation of grassland areas at both sierras. The impact of DOMÍNGUEZ, P.A. ( ) Efecto del cambio climático en la distribución del conejo endémico de México Romerolagus diazi. Bachelor’s thesis, such artificial creation of forest on the habitat of the volcano Universidad Nacional Autónoma de México, Mexico City, Mexico. rabbit should be evaluated. In addition, a long-term moni- EDGEL, R.J., PIERCE, J.L. & LARSON, R.T. () Pygmy rabbit toring programme for the volcano rabbit should be estab- (Brachylagus idahoensis) habitat selection: does sagebrush lished, to examine the effects of both the natural com- (Artemisia spp.) age influence selection? Western North American ponents of the habitat and human disturbances on the Naturalist, , –. dynamics of extinction and recolonization of local popu- FA, J.E., ROMERO, F.J. & LOPEZ-PANIAGUA,J.( ) Habitat use by parapatric rabbit in a Mexican high-altitude grassland system. lations of the volcano rabbit and other endemic species. Journal of Applied Ecology, , –. GALICIA,L.&GARCÍA-ROMERO,A.() Land use and land cover Acknowledgements We thank C. Osuna, F. Osuna and M. Urios- change in highland temperate forest in the Izta-Popo National Park, tegui for their support during fieldwork; The Rufford Foundation for central Mexico. Mountain Research and Development, , –. financial support (Rufford Small Grant 21227-1); and A. Smith and GREEN, J.S. & FLINDER, J.T. () Brachylagus idahoensis. one anonymous reviewer for their critiques. FO received a Graduate Mammalian Species, , –. Student Scholarship from CONACyT México (ID: 295207). This HUNTER,M.&CRESWELL,W.() Factors affecting the distribution manuscript is submitted in partial fulfilment of the requirements for and abundance of the Endangered volcano rabbit Romerolagus diazi the degree Doctorado en Ciencias of the graduate student programme on the Iztaccihuatl volcano, Mexico. Oryx, , –. at the Instituto de Ecologia AC, México. INEGI (INSTITUTO NACIONAL DE ESTADÍSTICA Y GEOGRAFÍA, MÉXICO)() Continuo de Elevaciones Mexicano . (CEM .). Author contributions Project design: all authors; logistic support: inegi.org.mx/app/geo/elevacionesmex [accessed June ]. AEM; fieldwork: FO; data analyses: FO, RG; writing: FO, RG, AEM. KOLASA,J.&LI, B.L. () Removing the confounding effect of habitat specialization reveals the stabilizing contribution of Conflicts of interest None. biodiversity to species variability. Proceedings of the Royal Society B, , –. Ethical standards This research abided by the Oryx guidelines on KREBS, C.J., GILBERT, B.S., BOUTIN,S.&BOONSTRA,R.() ethical standards. No animals were collected or handled, and site Estimation of snowshoe hare populations density from turd surveys had minimal impact on the habitat. transects. Canadian Journal of Zoology, , –. LEACH, K., MONTGOMERY, W.I. & REID,N.() Biogeography, macroecology and species’ traits mediate competitive interactions References in the order Lagomorpha. Mammal Review, , –. MÁRQUEZ, A., VERMA, S.P., ANGUITA, F., OYARZUN,R.&BRANDLE, ALVES, P.C., FERRAND,N.&HACKLÄNDER,K.() Lagomorph J.L. () Tectonics and volcanism of Sierra Chichinautzin: Biology: Evolution, Ecology and Conservation. Springer Verlag, extension at the front of the Central Trans-Mexican Volcanic Belt. Berlin Heidelberg, Germany. Journal of Volcanology and Geothermal Research, , –. ARREGOITIA, L.V., LEACH, K., REID,N.&FISHER, D.O. () MARTÍNEZ-GARCÍA, J.A., MENDOZA, G.D., ALCANTARA, J.L., Diversity, extinction, and threat status in lagomorphs. Ecography, TARANGO, L.A., SÁNCHEZ, T., RODRÍGUEZ,R.&HERNÁNDEZ, P.A. , –. () Composición de la dieta y capacidad nutricional de carga del BLANCO, S., CEBALLOS, G., GALINDO, C., MAAS, M., PATRÓN, R., hábitat del conejo de los volcanes (Romerolagus diazi) en México. PESCADOR,A.&SUÁREZ,A.() Ecología de la Estación Revista Chapingo, Serie Ciencias Forestales y del Ambiente, Experimental Zoquiapan. Cuadernos Universitarios, Universidad , –. Autónoma de Chapingo, Chapingo, Mexico. MONROY, M.R., PINO, M.J.M., LOZANO, G.M.A. & GARCÍA, F.A. CERVANTES, F.A. () Some predators of the zacatuche (Romerolagus () Estudio etnomastozoológico en el Corredor Biológico diazi). Journal of Mammalogy, , –. Chichinautzin. Sitientibus Série Ciéncias Biológicas, , –.
Oryx, Page 9 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368 10 F. Osuna et al.
MUELLER-DOMBOIS,D.&ELLENBERG,H.() Aims and Methods SCHMALZ, J.M., WACHOCKI, B., WRIGHT, M., ZEVELOFF, S.I. & of Vegetation Ecology. John Wiley, New York, USA. SKOPEC, M.M. () Habitat selection by the pygmy rabbit OSUNA, F., GONZALEZ, D., ESPINOSA DE LOS MONTEROS,A.& (Brachylagus idahoensis) in Northeastern Utah. Western North GUERREO, J.A. () Phylogeography of the volcano rabbit American Naturalist, , –. (Romerolagus diazi): the evolutionary history of a mountain specialist SEMARNAT (SECRETARIA DE MEDIO AMBIENTE Y RECURSOS moulded by the climatic-volcanism interaction in the central Mexican NATURALES)() Norma Oficial Mexicana highlands. Journal of Mammalian Evolution, , –. NOM--SEMARNAT-. Protección Ambiental-Especies PIERCE, J.E., LARSEN, R.T., FLINDERS, J.T. & WHITING, J.C. () Nativas de México de Flora y Fauna Silvestres-Categorías de Riesgo y Fragmentation of sagebrush communities: does an increase in habitat Especificaciones para su Inclusión, Exclusión o Cambio—Lista de edge impact pygmy rabbits? Animal Conservation, , –. Especies en Riesgo. Diario Oficial de la Federación a sección, PORTALES, G.L., REYES, P., RANGEL, H., VELAZQUEZ, A., MILLER,P., Mexico City, Mexico. ELLIS,S.&SMITH, A.T. () Taller Internacional para la SEMARNAT (SECRETARIA DE MEDIO AMBIENTE Y RECURSOS Conservacion de los Lagomorfos Mexicanos en Peligro de Extinción. NATURALES)() Reglamento Interior de la Secretaría de Medio IUCN/SSC Conservation Breeding Specialist Group, Apple Valley, USA. Ambiente y Recursos Naturales. Diario Oficial de la Federación, QGIS DEVELOPMENT TEAM () QGIS Geographic Information November . Mexico City, Mexico. System .. Open Source Geospatial Foundation Project. qgis.org TREJO,I.&DIRZO,R.() Deforestation of seasonally dry tropical [accessed June ]. forest: a national and local analysis in Mexico. Biological RDEVELOPMENT CORE TEAM () R: A Language and Environment Conservation, , –. for Statistical Computing. R Foundation for Statistical Computing, UNITED STATES DEPARTMENT OF STATE GEOGRAPHER () Google Vienna, Austria. Earth Pro. Google Earth. google.com/intl/es/earth [accessed RIZO-AGUILAR, A., GUERRERO, J.A., MONTOYA-LARA, A.M.P. & June ]. VALDESPINO,C.() Physiological stress in volcano rabbit URIOSTEGUI-VELARDE, J.M., GONZALEZ-ROMERO, A., PINEDA, E., Romerolagus diazi populations inhabiting contrasting zones at the REYNA-HURTADO, R., RIZO-AGUILAR,A.&GUERRERO, J.A. Corredor Biológico Chichinautzin, Mexico. Mammalian Biology, () Configuration of the volcano rabbit (Romerolagus diazi) , –. landscape in the Ajusco-Chichinautzin mountain range. Journal of RIZO-AGUILAR, A., GUERRERO, J.A., HIDALGO-MIHART, M.G. & Mammalogy, , –. GONZÁLEZ-ROMERO,A.() Relationship between the VELÁZQUEZ,A.&HEIL, G.W. () Habitat suitability study for the abundance of the Endangered volcano rabbit Romerolagus diazi and conservation of the volcano rabbit (Romerolagus diazi). Journal of vegetation structure in the Sierra Chichinautzin mountain range, Applied Ecology, , –. Mexico. Oryx, , –. VELÁZQUEZ,A.&GUERRERO, J.A. () Romerolagus diazi.In RIZO-AGUILAR, A., DELFIN-ALONSO, C., GONZALEZ-ROMERO,A.& The IUCN Red List of Threatened Species : e.TA. GUERRERO, J.A. () Distribution and density of the zacatuche dx.doi.org/./IUCN.UK.-.RLTS.TA.en rabbit (Romerolagus diazi) at the protected natural area ‘Corredor [accessed October ]. Biologico Chichinautzin’. Therya, , –. VELÁZQUEZ, A., ROMERO, F.J. & LÓPEZ-PANIAGUA,J.() Ecología RZEDOWSKI,J.() Vegetación de México. Limusa, Mexico City, y conservación del conejo zacatuche (Romerolagus diazi) y su hábitat. Mexico. Fondo de Cultura Económica, Mexico City, Mexico. SANTILLI, F., BAGLIACCA,M.&PACI,G.() Density and habitat ZHAO, C., YUE, B., RAN, J., MOERMOND, T., HOU, N., YANG,X.&GU, use of sympatric brown hares and European rabbits in a X. () Relationship between human disturbance and Endangered Mediterranean farmland area of Tuscany (central Italy). giant panda Ailuropoda melanoleuca habitat use in the Daxiangling Ethology Ecology & Evolution, , –. Mountains. Oryx, , –.
Oryx, Page 10 of 10 © The Author(s), 2021. Published by Cambridge University Press on behalf of Fauna & Flora International. doi:10.1017/S0030605320000368 Downloaded from https://www.cambridge.org/core. IP address: 170.106.202.126, on 28 Sep 2021 at 19:39:59, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/S0030605320000368