Species distribution modeling for the section Xerobia Bunge of the genus Oxytropis DC. on the territory of Central Asia under past and future climate change

Abstract

Modern botanical studies revealing patterns of plant species distribution are based on analysis of big datasets. Despite publishing many maps of diversity and species richness on the global scale and for huge biogeographic regions of the world, the territories of Northern and Central Asia remain poorly studied. We elaborated a special database, including distribution of 19 Oxytropis species of the section Xerobia with 1353 localitites (See Fig. 1). For all species, we analyzed their whole distribution range, including data from Kazakhstan, Mongolia and China. Species distribution was detected according to the main foreign and Russian herbaria, online databases, field data and relevés. Species distribution modeling was performed using Maxent 3.3.3k with MIROC-ESM model in resolution of 2.5 arc-minutes. 19 BIOCLIM and 18 ENVIREM variables were analyzed. Past climate change was evaluated using ENVIREM variables for the Mid-Holocene (ca. 6.000 yr. BP) and the Last Glacial Maximum (ca. 22.000 yr. BP). Future distribution modeling was carried out basing on different climatic scenarios, according to IPCC AR5: RCP8.5, RCP2.6 and RCP6.0. Species distribution from the section Xerobia mostly occupied the territory of Central Asia (See Fig. 1). Few species, such as Oxytropis grandiflora (Pall.) DC. and O. leptophylla (Pall.) DC., were mostly found in the western part of Xerobia section distribution on the territory of Zabaykal’skiy region of Russia, Eastern province of Mongolia and Inner Mongolia province of China. The most part of Xerobia species have isolated distribution and often occupy specific habitats. In such case, using SDM with only bioclimatic variables for local endemic species is pointless. So, we chose species Oxytropis ampullata (Pall.) Pers. (See Fig. 2) with Central Asian distribution and O. grandiflora with Manchuro-Dahurian distribution for modeling (See Fig. 3A). The selected species differ in their ecology: O. ampullata is a mountainous species, whereas most habitats for O. grandiflora are river valleys and mid-mountainous regions. Our analysis showed that ENVIREM variables provide more correct modeling results than BIOCLIM variables (See Fig. 2). Predictive maps on the basis of BIOCLIM variables showed wide potential distribution for O. ampullata, which does not correspond well to the species ecology. The main habitats for this species are such mountainous regions as the Khangai mountains, the Russian and the Mongolian Altai mountains, the Dzhungarian mountains, and the Tarbagatai ridge. Additionally, modeling showed potential distribution for the species in the Selenga river valley. Modern distribution of O. grandiflora was studied quite well; suitable habitats with new localities for the species can be found in the Khentii mountains (See Fig. 3A). The determinants for O. ampullata are mean annual temperature, isothermality and potential evapotranspiration (PET) of the driest and coldest quarter (See Table 1). PET parameters in the driest and the coldest time of the year have the key meaning because in arid conditions plants receive the main portion of moisture in the colder period when the evaporation is not intense, also it is important to conserve the moisture during the dry season. Distribution of O. grandiflora is limited by temperature and precipitation seasonality, temperature annual range, PET seasonality, and PET of the driest quarter (See Table 1). Determinants for the species with Central-Asian distribution O. ampullata are connected with temperature variables, whereas for Manchuro-Daurian species O. grandiflora precipitation matters (See Table 1 and Fig. 3A). The key factors for modern distribution of the studied Xerobia species are mean monthly potential evapotranspiration of the driest quarter, continentality index and climatic moisture index (See Fig. 3B). All these variables were determinants for the mid-Holocene and the Last Glacial Maximum (See Table 2 and Fig. 4), which might give evidence of relatively stable environmental conditions in the studied region. Central Asia has not been severely affected by glaciation as more northern latitudes and climate conditions on that territory were relatively stable during a long period. Modeling for the past climate showed a wider distribution for Xerobia species in the north-west during the Last Glacial Maximum and future shrinking during the Mid-Holocene till modern time (See Fig. 4). The north-eastern territories, such as Zabaykal’skiy region of Russia and, partially, the central part of Siberia, are characterized by a wider distribution under modern climate conditions. Species habitats of that territory are mostly confined with mountains. It is consistent with previous studies that described Southern Siberia as one of the centers of speciation for the genus Oxytropis. This region has now high Oxytropis species richness with a great number of endemics.Predictive maps for different climate scenarios reveal insignificant changes in distribution of the section Xerobia, even for the maximum climate warming (RCP8.5 scenario) (See Fig.5). Under predicted climate change, potential habitats in the southwest and in the north-east of Xerobia distribution, as well as a slight shrinking in the south-east can be observed in the future.