Diversity and Distribution Patterns of Geometrid Moths (Geometridae, Lepidoptera) in Mongolia

Geometrids are a species-rich group of moths that serve as reliable indicators for environmental changes. Little is known about the Mongolian moth fauna, and there is no comprehensive review of species richness, diversity, and distribution patterns of geometrid moths in the country. Our study aims to review the existing knowledge on geometrid moths in Mongolia. We compiled geometrid moth records from published scientific papers, our own research, and from the Global Biodiversity Information Facility (GBIF) to produce a checklist of geometrid moths of Mongolia. Additionally, we analyzed spatial patterns, species richness, and diversity of geometrid moths within 14 ecoregions of Mongolia and evaluated environmental variables for their distribution. In total, we compiled 1973-point records of 388 geometrid species. The most species-rich ecoregion in Mongolia was Daurian Forest Steppe with 142 species. Annual precipitation and maximum temperature of the warmest month were the most important environmental variables that correlated with NMDS axes in an analysis of geometrid assemblages of different ecoregions in Mongolia.


Introduction
Regarded as disturbing pests or less charismatic than butterflies, moths are nevertheless creatures with an important role in the ecosystem and the potential to serve as environmental indicators [1][2][3][4]. Moths are globally distributed and it is estimated that more than 130,000 described species exist [5], far more than the more conspicuous and mostly diurnal butterflies with ca. 20,000 species. Many moths are pollinators, but due to their nocturnal activity they are not well studied [6]. In a recent review from the current literature, Hahn and Brühl reported that in Europe and North America there are 227 moth-plant interactions with 129 moth species involved [6]. Geometrid moths (Geometridae), constituting one of the biggest families of Lepidoptera, are a species-rich and easily recognizable family that have served as indicators for environmental changes in many previous studies [7][8][9][10]. These groups also appear to be effective at colonizing habitats after natural or anthropogenic disturbances [11]. There are approximately 24,000 described species of Geometridae worldwide [12]. Although Mongolia is one of the largest countries (rank 19th in size) on Earth, little is known about its moth fauna, and there is no comprehensive review of species richness, diversity, and distribution patterns of geometrid moths in the country. A few researchers attempted to summarize information to mainly confirm this lack of information [13].
Mongolia is a country that encompasses landscapes with a high variety of climatic and geographic features with forest in the north, high mountains in the west, desert in the south, and steppes in the steppes in the eastern and central parts of Mongolia [14,15]. Altogether, it comprises 16 ecoregions [16] (Figure 1). Ecosystems change along a latitudinal gradient from forest in the north, over steppe and semi-desert to desert in the south [17]. In most areas of the country, livestock herding is a dominant land-use practice, and due to overgrazing, some pasture lands have recently been degraded [18]. With recent discoveries of various mineral resources, mining has become not only the main economic sector, but also the major reason for environmental disturbance in Mongolia. Together with climate change, it is the major driver for habitat loss and environmental changes [14,19]. As a result of these anthropogenic changes, many species are disappearing, but there is little information about which species are at greatest risk of becoming extinct, especially for the less studied taxa.
In order to monitor diversity loss and gain, and to further study the influence of environmental disturbance and climate change on geometrid moths in Mongolia, we need an up-to-date dataset that mirrors the current state of knowledge and that includes all species already recorded. Given this knowledge gap, this study aims to review, summarize, and evaluate the existing knowledge on geometrid moths in Mongolia. It will provide a baseline for further studies, as well as define research priorities in the field. In this study, we aim to: (1) provide a checklist of geometrid moths of Mongolia, setting a baseline for future studies, (2) analyze distribution patterns and species richness and diversity of geometrid moths within ecoregions of Mongolia, and (3) analyze which environmental variables are most important in determining their distribution. We are aware that all results can only give a provisional status due to the data situation, especially the results for Objectives 2 and 3 can only be given with caution; however, our detailed review of the current data will help to define the needs for further research more efficiently.

Study Review
Information on the species composition of Macrolepidoptera of Mongolia began to accumulate from the end of the nineteenth century, as a result of the works of collectors such as Fritz Dörries, Hauberhauer and Leder, and others. Otto Staudinger [20] published the first paper on the collection of Fritz Dörries, who made a trip in 1879 to Khentii Mountains to collect Lepidoptera. This resulted in data on the location of 75 species of geometrids in central and western parts of Mongolia [20]. Later, Staudinger published several papers and books on the fauna of Palaearctic Lepidoptera which included some geometrid species from Mongolia [21][22][23]. In 1964, a Mongolian-German expedition conducted a biological survey, as a result of the expedition 214 Lepidopteran exemplars were sampled. Burchard Alberti later published the results on Lepidoptera and nine geometrid species were listed in the paper [24]. Likewise, Joseph Moucha listed four geometrid species from a Mongolian-Czech entomological-botanical expedition, which was conducted around 1960 [25]. Grigory Grum-Grshimailo found three geometrid species from Selenge Aimag in the collection of M.I. Molleson [26]. Alexander Mikhailovich Djakonov [27,28] recorded a new occurrence of Horisme scosiata and described one new species Scotopteryx transbaicalica from the family of Geometridae based on old material of Staudinger. Other researchers such as Karl Dietze [29], Eugen Wehrli [30], and Fritz Heydemann [31] also described new species. In the fourth volume and its supplementary of "Die Gross-Schmetterlinge der Erde. Die Spanner des Palaearktischen Faunengebietes" series edited by Adalbert Seitz, 34 geometrid species were listed for Mongolia [32,33].
The most important contribution to the collection and study of Mongolian geometrid moths were made by Russian and Soviet expeditions led by Pyotr Kuzmich Kozlov and later by Soviet-Mongolian expeditions [34][35][36]. During the survey of Soviet-Mongolian expeditions, Jaan Viidalepp recorded a total of 201 geometrid species.Viidalepp later in 1999 compiled a checklist of geometrid moths of the former U.S.S.R and in this monograph 210 species were included for Mongolia [37]. Particularly rich and diverse material on Lepidoptera (41,000 specimens) were collected by the Hungarian expeditions conducted by Zoltán Kaszab, who made six entomological collecting trips along latitudinal and longitudinal gradients in Mongolia, between 1963 and 1968. András Vojnits published several papers based on the Kaszab's collections dedicated to subfamilies of Geometridae in the period between 1974 and 1979. He recorded 177 species from the whole collection, described 39 species new to the fauna of Mongolia and four species new to science [38][39][40][41][42][43][44]. Malcolm J. Scoble [45] presented 66 taxa from Mongolia.
Other researchers also contributed to the study of Mongolian geometrid moths. For instance, Gantigmaa Ch. and coworkers recorded 90 species in the West Khentii of Northern Mongolia [46]. In the book "Biodiversity of Sokhondinsky Reserve", 29 geometrid species from Mongolia have been included [47]. Beljaev and Vasilenko [48] noted 29 species of geometrid moths in Mongolia. Vasilenko and colleagues [49][50][51] recorded eight species and described one new species Rhodostrophia ustyuzhanini in Western Mongolia. In 2012 and 2013, we collected 70 geometrid species from central and northern parts of Mongolia [4]. Mironov and Glasworthy [52] reported 57 species with two species (Eupithecia ankini, Eupithecia munguata) new to science and 12 species new to the fauna of Mongolia. Erlacher et al., studied six geometrid species from Mongolia and described one new species Charissa beljaevi [53][54][55]. In 2019, Makhov and Beljaev [56] studied the geometrid moths of the Baikal Region and recorded 14 species from Mongolia. In six volumes of "The Geometrid Moths of Europe", 117 moth species are listed from Mongolia. We validated our species checklist with these volumes [57][58][59][60][61][62].

Materials and Methods
We compiled geometrid moth records from published scientific papers, from our work [63] (all sample identifications were double checked by curator T. Enkhbayar, Department of Biology, National University of Mongolia), from the collections of the Siberian Zoological Museum (curator -S.V.Vasilenko) [64], and also from the Global Biodiversity Information Facility (GBIF) [65]. Lastly, we checked the "Revised, annotated systematic checklist of the Geometridae of Europe and adjacent Diversity 2020, 12, 186 4 of 26 areas, Vols 1-6" [62]. From the Museum collections we could only get country-level information, not the exact location. From GBIF data, we included 380 records into our species list [65]. Fourteen specimens of six species were found in the public data of The Barcode of Life Data System (Bold System) [66].
We  (Table S1). Of these records, 87 species were missing information on exact locations, these 87 species are used to estimate species richness and listed in the species checklist but are excluded from other analysis. We georeferenced species locations from literature and generated coordinates of each location with Google Earth [67]. After that we cross-checked each species name in "The Global Lepidoptera Names Index" [68]. Moreover, experts on geometrid moths such as Axel Hausmann, Jaan Viidalepp, Gunnar Brehm, Sven Erlacher, and Pasi Sihvonen validated most species of our checklist and provided further literatures.
In the next step we used the sampled data in order to estimate true species richness, to evaluate the distribution of species within Mongolia, and to identify regions that have been undersampled so far by species rarefaction. For these reasons, we transformed all species locations into 2 • × 2 • grid cells, resulting in 51 grid cells inhabited by 301 species. Of 301 species, 121 were unique species occuring only once within 51 grids. To estimate species richness we applied Good Turing Theory, which uses unique species for estimation [69]. We used the application SuperDuplicates (https://chao.shinyapps.io/SuperDuplicates/) for the estimation with the following setting: Data type: incidence data; Number of observed species (SOBs): 388; Number of uniques (Q1): 208 (we combined the 121 unique species with the former mentioned 87 species without locations).
Further we calculated rarefaction curves for single ecoregions to assess collection quality in different areas of Mongolia. Four ecoregions (Altai Alpine Meadow and Tundra, Dzungarian Basin Semi-Desert, Khangai Mountains Alpine Meadow and Sayan Alpine Meadows, and Tundra) were strongly under sampled, having species richness below 15, thus we excluded them from the analysis to avoid misleading interpretation.
To estimate the rarefaction curve across grid cells and ecoregions, we calculated interpolation and extrapolation of species richness using the 'iNEXT' package: Interpolation and extrapolation for species richness in R [70,71] with 0.95 confidence interval and prepared the rarefaction plots with 'devtools' package [72] and ggiNEXT function of 'ggplot2 package [73].
We performed Non-Metric Multidimensional Scaling Analysis (NMDS) to check the dissimilarity of geometrid species composition between ecoregions based on the zero-adjusted Bray-Curtis dissimilarity measure using 'phytomosaic/ecole' and 'vegan' package [74][75][76]. For estimation of pairwise similarities between ecoregions, we calculated the estimated abundance based Soerensen Index by abundance data using online program SpadeR [77]. We preferred Soerensen Index over Jaccard Index, while the result was a little bit higher than Jaccard. This estimated abundance based index can detect unseen shared species and is appropriate to evaluate beta diversity of samples under sampling bias [78].
We used 19 Bioclim data with 30 arc seconds resolution as climatic variables for the region [79]. We extracted these variables for the fourteen ecoregions. Ecoregion GIS data for Mongolia were downloaded from The Nature Conservancy (TNC) [80]. In two ecoregions no geometrid moths were found, namely, Khangai Mountains Conifer Forests and Sayan Intermontane Steppe ( Figure 1). We thus excluded these ecoregions from the further analysis. To check for strong linear dependencies among explanatory variables we computed the variance inflation factor (VIF) for each variable in R package 'vegan'. We excluded variables with VIF values higher than 10 [81] (Table 1). We chose the most significant environmental variables with forward selection method by using vegan's 'ordistep' function [81]. Variables selected by forward selection method were fitted into the ordination plot using vegan's 'entfit' function.
All analysis were performed in R [82] and most graphs were made with package 'ggplot2 [73].

Results
Altogether, we recorded 388 geometrid species of six subfamilies: Archiearinae, Desmobathrinae, Ennominae, Geometrinae, Larentiinae, and Sterrhinae (Appendix A Table A1). The most species-rich subfamily was Larentiinae with 203 species, while we recorded only one species in the subfamily Desmobathrinae. For 301 species with exact location data (Table S1), we recorded species richness within 2 • × 2 • grid cells in whole Mongolia ( Figure 2). Altogether, we recorded 388 geometrid species of six subfamilies: Archiearinae, Desmobathrinae, Ennominae, Geometrinae, Larentiinae, and Sterrhinae (Appendix Table 1). The most species-rich subfamily was Larentiinae with 203 species, while we recorded only one species in the subfamily Desmobathrinae. For 301 species with exact location data (Table  S1), we recorded species richness within 2° × 2° grid cells in whole Mongolia ( Figure 2).   In the next step we used the fourteen Mongolian ecoregions ( Figure 1) to investigate the distribution of the sampled geometrid species in more detail. The most species-rich ecoregion was Daurian Forest Steppe with 142 species, while Khangai Mountains Alpine Meadow was the lowest in species richness with only three species of geometrid moths ( Figure 4). One species (Rhodostrophia jacularia) occurred in 10 ecoregions, there were five further generalist species (Euphyia unangulata, Eupithecia nephelata, Scopula albiceraria, Scopula beckeraria) that occurred in eight to nine ecoregions. In contrast, 126 species were recorded only in one ecoregion. Four ecoregions were clearly under-sampled (Altai Alpine Meadow and Tundra, Dzungarian Basin Semi-Desert, Khangai Mountains Alpine Meadow, Sayan Alpine Meadows and Tundra) thus to avoid misleading interpretation, we excluded those ecoregions from further analysis.    Table 2. Pairwise estimates of similarity between ecoregions with online tool Spade [69]. Shown is the estimated abundance-based Sorensen Index. Colors refer to the map in Figure 1. Ecoregion abbreviations as in Figure 4. Highest and lowest values in bold. An NMDS ordination biplot (stress = 0.05) shows two separate groups of geometrid species communities within ecoregions ( Figure 6). Altai Montane Forest and Forest Steppe, Alashan Plateau Semi-Desert, Eastern Gobi Desert Steppe, Gobi Lakes Valley Desert Steppe, and Great Lakes Basin Desert Steppe are clustered in the first group, Sayan Montane Coniferous Forests, Mongolian-Manchurian Grassland, Daurian Forest Steppe, Selenge-Orkhon Forest Steppe, and Trans-Baikal Coniferous Forests are grouped in the second group. Precipitation was positively correlated with NMDS1, while temperature was positively correlated with NMDS2, both correlations were highly significant (p < 0.01). Number of records was positively correlated with both axes but was not significant (Table 3).   For assessment of beta-diversity, we calculated estimates of the abundance-based Sorensen Index between ecoregions (Table 2). We excluded ecoregions with fewer than 20 species to avoid sampling bias in similarity analysis. The highest pairwise estimated Sorensen Similarity Index was between Eastern Gobi Desert Steppe and Gobi Lakes Valley Desert Steppe (β s = 0.942), while the lowest were between Trans-Baikal Coniferous Forests and both of Gobi Lakes Valley Desert Steppe, Great Lakes Basin Desert Steppe (β s = 0.076). Table 2. Pairwise estimates of similarity between ecoregions with online tool Spade [69]. Shown is the estimated abundance-based Sorensen Index. Colors refer to the map in Figure 1. Ecoregion abbreviations as in Figure 4. Highest and lowest values in bold. An NMDS ordination biplot (stress = 0.05) shows two separate groups of geometrid species communities within ecoregions ( Figure 6). Altai Montane Forest and Forest Steppe, Alashan Plateau Semi-Desert, Eastern Gobi Desert Steppe, Gobi Lakes Valley Desert Steppe, and Great Lakes Basin Desert Steppe are clustered in the first group, Sayan Montane Coniferous Forests, Mongolian-Manchurian Grassland, Daurian Forest Steppe, Selenge-Orkhon Forest Steppe, and Trans-Baikal Coniferous Forests are grouped in the second group. Precipitation was positively correlated with NMDS1, while temperature was positively correlated with NMDS2, both correlations were highly significant (p < 0.01). Number of records was positively correlated with both axes but was not significant (Table 3).

Discussion
In this study, we compiled a geometrid species checklist for Mongolia, examined species richness and diversity of geometrid communities among ecoregions. In addition, we investigated which environmental variables impact the distribution of geometrid moths. Compiling a species checklist on geometrid moths from a variety of sources published since 1892 was quite challenging, as names of species and locations were changing over the years, while sample efforts in different studies and areas differed considerably. Despite all our efforts we may not have included all species recorded in Mongolia in our list.
In total, we found 1973 records of 388 geometrid species of six subfamilies, but these records were not evenly sampled. The sample-based interpolation and extrapolation curve of gridded sample was not asymptotic, indicating that our records do not represent the whole potential geometrid fauna in Mongolia (Figure 3). Species richness for whole Mongolia was estimated as 663.19 species with Good-Turing theory and this estimated species richness was nearly double the observed species richness. These results confirm the rarefaction analysis and show that our inventory of geometrid moths in Mongolia is still incomplete, with less than 60% of the estimated species being recorded. The fact that countrywide diversity was highest in the grid cell of the capital draws further attention towards an obvious sampling bias with undersampling for the rest of the country. Moreover, we expect to find species of two other subfamilies, Orthostixinae and Alsophilinae in Mongolia. Species of these subfamilies were recorded in adjacent areas, such as in Kazakhstan and in China [37]. However, according to Müller et al. Alsophilinae is transferred to Ennominae, while the subfamily status of Orthostixinae is still not clear [62].
Given the huge size of Mongolia the estimated richness of 663 geometrid species for the whole country seems to be not high. But we wanted to compare the species richness of Mongolia with species richness of other countries similar in size. Norway + Sweden + Finland (1,173,940 km 2 ) together are similar in size to Mongolia (1,564,000 km 2 ). Altogether, for these countries, 341 geometrid species are recorded [84]. If we compare observed species richness (388) of Mongolia with the richness of those countries, it is almost similar; if we compare estimated species richness (663), it is almost double.
However, Scandinavia is an area at high latitudes, with harsh climate, not really suited for an ectotherm group like moths. Further south, Iberian Peninsula and Balearic Islands together, have 589 geometrid species (According to a personal information of Javier Gastón, one of the authors of the paper, due to scientific efforts the total number of Geometridae recorded on Iberian Peninsula and the Balearic Islands is now 605 species.) [85] and their areas (596,740 km 2 + 4564 km 2 ) are almost three times smaller than the landlocked area of Mongolia, which is situated at higher latitude. Comparisons between distant countries are always somewhat lacking, but no figures on geometrid species richness are available for the countries in Inner Asia (e.g., Kazakhstan).
The most frequently recorded species, which occurred in 10 ecoregions of Mongolia, was Rhodostrophia jacularia, an inhabitant of steppe and semi-desert [34,86]. Sihvonen and Nupponen [87] studied female wing shape of this species, but we could not find other studies related to the biology of this species.
Most records were found in Daurian Forest Steppe, Selenge-Orkhon Forest Steppe, and Mongolian-Manchurian Grassland. For many ecoregions, rarefaction curves were not asymptotic, thus revealing that sampling there was incomplete. Two ecoregions have no geometrid moth records at all and were thus excluded from analysis, namely Khangai Mountains Conifer Forests and Sayan Intermontane Steppe. The less studied areas comprise higher altitude areas from central Mongolia, as well as border regions. Sampling in these ecoregions, many of them with high habitat heterogeneity, will certainly expand our checklist.
To assess beta diversity among these unevenly sampled groups we used an estimator for Soerenson similarity that includes unseen species in the calculation [70]. The results, on the one hand, reflect the high habitat heterogeneity of Mongolia, with is steep ecological north-south gradient and the diverse biomes of the country that promote high beta diversity (Table 1). On the other hand, it proved that ecoregions that include similar biomes had higher similarity of moth communities, a result corroborated by NMDS. The most similar ecoregions were Eastern Gobi Desert Steppe and Gobi Lakes Valley Desert Steppe that adjoin each other (β s = 0.942).
In NMDS, ecoregions were grouped in two big groups. The first group included Alashan Plateau Semi-Desert, Eastern Gobi Desert Steppe, Gobi Lakes Valley Desert Steppe, Great Lakes Basin Desert Steppe and Altai Montane Forest and Forest Steppe, while in the second group there were Daurian Forest Steppe, Mongolian-Manchurian Grassland, Sayan Montane Coniferous Forests, Selenge-Orkhon Forest Steppe, and Trans-Baikal Coniferous Forests. The geographically nearest ecoregions were grouped together, and also the ecoregions included in the same group belonged to mostly same biome type (Table 1). The first group comprised mostly Deserts and Xeric Shrublands except Altai Montane Forest and Forest Steppe, while three ecoregions of the second group belonged to Temperate Grasslands, Savannas and Shrublands.
Environmental variables that shaped species distribution were nominated by forward selection in NMDS and included annual precipitation and maximum temperature of warmest quarter. Number of records was also selected as variable, but only temperature and precipitation were significant in NMDS, thus corroborating the general robustness of our analysis, which was less influenced by sample effort. The aforementioned groups of ecoregions in NMDS differ along the precipitation gradient and within groups in temperature, e.g., the montane forests regions of both groups have lower values of NMDS2.
In a study on Borneo, geometrid moths showed a similar relationship with precipitation and temperature [88]. Temperature has also been a major impact on geometrid species distribution in the Andes [89]. Moreover, habitat disturbance played a big role in shaping the geometrid moth ensemble in northern Borneo [90]. Similarly, grazing proved to be a factor influencing community pattern in Mongolian moths [4]. Temperature, rainfall and habitat disturbance are impacted by climate change and anthropgenic influence, so we expect future changes within the Mongolian geometrid communities. The species list we present here can be a tool helping to monitor these changes.
Finally, we have to admit that our study has a few weaknesses. We compiled records only from literature (we apologize if we missed any) due to limited time and funding. A total of 87 of the 388 species in our checklist are still missing an exact location. This information may be available in the museum collections pinned to the respective specimens. A detailed research in museums would have certainly brought more records and species. In addition, all our records were not systematically collected, which might affect the statistical analysis. The mere fact that data were sampled over a long period of time in different research projects, with different ways of sampling certainly impacts the value of a statistical analysis. For example, in our field study [4], we used UV light, but in other studies normal light bulbs were used, sometimes even moths have even been collected during day time. Together with the general problem of undersampling, these points hamper a more detailed analysis of the Mongolian geometrid communities at the present time.
Nevertheless, due to our study, future directions of research on Mongolian Geometridae have become more clear: geometrid moths are really under-studied in Mongolia. We found two unsampled and four extremely under-sampled ecoregions and for all ecoregions expected species numbers were higher than recorded ones. So, we expect to find many more amazing moth species in future collections in the respective regions.

Conclusions
In total, 1973 records of 388 species were recorded, but we also expect that many more species will be recorded in the future in more elaborated sampling designs, especially from locations of southern, eastern and western Mongolia. Despite the fact that our compiled data is not good enough to analyze the distribution and diversity pattern in full detail, our study could reveal the knowledge gaps and undersampled areas, provide a first estimate of the approximate species number in whole Mongolia (n = 663), visualize the currently recorded distribution and diversity pattern of geometrid moths of Mongolia and evaluate the main environmental factors that shape the communities.

Acknowledgments:
We thank Reinhold Stahlmann for his support in GIS techniques and are grateful to our colleagues from the Department of Biogeography, University of Bayreuth, for their helpful comments on earlier drafts of the manuscript. We would like to express our deepest appreciation to anonymous referees for their detailed comments and useful suggestions, as these comments and suggestions led us to a significant improvement of the work and opened helpful contacts. We are grateful for Gunnar Brehm, Jena, for his kind assistance in providing important literature. We thank Javier Gastón, who provided us with updated information on the geometrid checklist of Iberian Peninsula. Finally, we deeply thank Axel Hausmann, Jaan Viidalepp, Sven Erlacher and Pasi Sihvonen for the review of the species checklist, their advise and provision of literatures.

Conflicts of Interest:
The authors declare no conflict of interest.