Changes in Spatial Distribution of Bryophytes under CMIP6 Future Projections on the Qinghai-Tibet Plateau

Bryophytes play important roles in ecosystem due to their extensive geographical coverage on the Qinghai-Tibetan Plateau (QTP). While there are few studies attributing the potential distribution and landscape changes on the QTP in response to climate change. Based on climate data averaged of nine global climate models (GCMs) for shared socio-economic pathways SSP2-4.5 under current (the years 1970–2000) and future climate scenarios (the years 2021–2040, 2041–2060, 2061–2080, 2081–2100), and other environmental variables, this study has applied the maximum entropy (MaxEnt) model to assess the potential impact of climate change on the distribution of Bryophytes on the QTP. The key environmental factors which determined Bryophytes’s habitats and range shifts were also examined. The results showed that Bryophytes occupied about 9.12 × 10 5 km 2 (35.43% of total QTP) at present, mainly accumulating in non-permafrost regions of southeast (SE) QTP. Niche suitability of the Bryophytes was dominated by soil moisture, ultraviolet-B radiation seasonality, temperature seasonality and precipitation of the coldest quarter. The occupied habitats of Bryophytes under future climate scenarios generally increased migrating towards Midwest and relatively higher elevation regions of QTP, where dedicated overall surface air warming and moistening, solar dimming. Additionally, the confusion matrix showed that most parts of the gained occupied habitats under future climate scenarios were low suitable habitats, and small parts for high suitable habitats, however reduced for the medium suitable habitats.


Introduction
The Qinghai-Tibetan Plateau (QTP), also called as "The Third Pole of the world", is particularly vulnerable to climate change (IPCC 2019; Pepin et al., 2015;Thompson et al., 2018). Signi cant changes gleaned from long-term in-situ and satellite observations, and modeling results all have shown to happened, including vegetation phenology, treeline, vegetation greening, permafrost degradation, solar radiation and vegetation feedback with the climate warming over the QTP Yao et al., 2019;Kang et al., 2019).
Bryophytes are widely distributed on the QTP (Gao et al., 2016). Thus, understanding the spatial distribution patterns and range shifts trends, and the key environmental factors, is helpful for protecting QTP ecosystems diversity, monitoring vegetation and climate changes, and guiding future eld surveys of new Bryophytes species' occurrences, especially in unexplored areas. However, we largely overlooked the vast areas distribution of Bryophytes and the response of the spatial extent of their suitable habitats to climate change (Wu et al., 2002b;Bao 2006, Ingerpuu et al., 2019). Both researches and the red list of endangered Bryophytes in China approved that global warming will lead to diversity losses, or even endangered for Bryophytes of QTP (Cao et al., 2006;He et al., 2016). But Wu et al (2001) issued that despite the fact that a small part of Bryophytes disappeared with the uplift of QTP, in general, warming climate prompted the development and geographical distribution range in Hengduan mountains during past decades, although vast geophysical surveys and control experimental con gurations about Bryophytes had been conducted (Wu et al., 2003a), but these ways are extremely costly and only available over relatively small areas. The maximum entropy model (Maxent), an numerical tool that only combine limited species presence (or occurrence) data and environmental variables (Elith et al., 2006;Phillips et al., 2009), has come into particularly common used species distribution models (SDMs) ) to gain species ecological and distributional evolutionary insights under changing climate (Muir et al., 2015;You et al., 2018;Guo et al., 2019). Maxent modeling was also considered as a useful method for mapping geographical distribution and response to climate change for Bryophytes on a large scale. For example, Sérgio et al. (2007) compared three different approaches: genetic algorithm for rule-set production (GARP), maximum entropy (Maxent) and ecological niche factor analysis (ENFA) to modeling the potential distribution of Bryophytes, and the accuracy of Maxent dedicate the best; Désamoré et al. (2012) investigated the distribution and genetic diversity of Bryophytes under past and present climates conditions; Skowronek et al. (2018) mapped the distribution of Bryophytes combing with imaging spectroscopy data in Germany and Belgium; Song (2015) predicted the potential distribution of Pottiaceae in Tibet and explored the key ecological factors. However there are no reports to forecast the potential geographical distribution and range shifts of Bryophytes on the whole QTP, especially in clarifying the change of habitat suitability with climate change.
The objectives of this work were: (1) to explore the key environmental factors affecting the habitat suitability of Bryophytes; (2) to delineate the potential distribution of Bryophytes under current and future climatic scenarios; and (3) to identify the differential effects of climate warming on the potential distribution and habitat suitability of Bryophytes.

Study regions and occurrence records
The occurrences of Bryophytes were shown in Fig. 1, which were collected from the Global Biodiversity Information Facility (GBIF, www.gbif.org/). The occurrences for Bryophytes in GBIF were from of ten datasets, such as EOD-eBird Observation Dataset, plant Specimens from PE Herbarium in China and Chinese Institute of Biology etc. All the occurrences of Bryophytes had passed strict quality preprocesses by deleting and ltering spatially to ensure that there no duplicate point within 10 km × 10 km before Maxent was begun. Finally, a total of 250 samples, representing all known Bryophytes natural habitats on the QTP, saved as .csv le format, were generated for further modeling.
Next, all above variables were resampling to a general spatial resolution of 30 s (ca. 1 km 2 ). Then we utilized Spearman's rank correlation to remove high spatially related bioclimatic variables based on previous reports of the factors potentially affecting Bryophytes, to avoid model over-tting lead by multicollinearity of variables, (Graham, 2008), with which highly collinear variables were identi ed, i.e., r> |0.75| (Suppl . Table 1). Finally, 16 bioclimatic variables (Table 1) were used to model the habitat distribution modeling. The performance of MaxEnt was evaluated by Area Under the receiver operating characteristics Curve (AUC) value (Pearson et al., 2006). AUC value ranged from 0.5 (random) to 1.0 (perfect discrimination) (Swets, 1988;Weber, 2011).

Land-Cover Transition Matrix
A land-cover transition matrix was generated in ArcGIS 10.0 software to re ect the changes of four suitable types of Bryophytes from one climate scenario to another (Wan et al., 2015).

Model performance
We determined AUC values for checking the model performance of all climate scenarios. The simulated results showed that the average AUC for the replicate runs of Maxent model under different climate scenarios was 0.9608 mean ± 0.038 SD ( Table 2). The results suggested that the performance of Maxent model was excellent.
The occupied habitats under the current climatic conditions mainly distributed in non-permafrost region of southeast (SE) QTP (Song et al., 2015;Zou et al., 2016) (Fig. 2). According to the classi ed results of HSI, the total occupied habitat reached to 9.12 × 10 5 km 2 (covered 35.43 % of total QTP), among which 18.46% was LSH; next was MSH (13.21%), while HSH accounted for the smallest percentage (3.75%). Geographically, the larger proportion of occupied habitats distributed in three rivers valley and the Yalu Tsangpo River basin. Moist condition and low UV-B radiation environment of these areas provided a suitable ecological corridor for Bryophytes (Martnez-Abaigar et al., 2003;Bartels et al., 2018).

Impact of climate change on Bryophytes
Variation of occupied areas of Bryophytes in four future time periods under SSP2-4.5 climate scenario was in accordance with total vegetation greening trend in QTP ecosystem with warming climate (Shen et al., 2016) (Fig. 3 and Fig. 4). With climate warming, degradation regions of permafrost and glacier in the central and western QTP would gradually become suitable for the growth of Bryophytes, and the potential occupied habitat of Bryophytes would migrate toward the Midwest of QTP (Fig. 3) Note that the shifts ratios of different habitat types changed inconsistently (Fig. 4). Ratios of LSH and HSH showed increasing trends, however fall in MSH. Concluding from the transition matrix results of four suitable habitat types (Table 3), the gained area of LSH mainly came from NSH and MSH, and increased area of HSH was from MSH. Totally, majority of lost MSH was converted to LSH, yet a small part to HSH. However, the area of the former was much larger than the latter, which was also consistent with the overall increase in the potential distribution of Bryophytes.  2021-2040, 2021-2040 to 2041-2060, 2041-2060 to 2061-2080 and 2061-2080 to 2081-2100 (unit: ×10 4  Both gained and lost occupied habitats of Bryophytes existed on the QTP as the temperature continues to rise. It could be inferred from Fig.5 that range shifts of Bryophytes correlated to elevations on the QTP, and gained habitats mainly migrate toward relatively higher elevations, while lost habitats happened at relatively low elevations.

Discussions
As an important bio-indication to climate monitoring, ecosystem biodiversity conservation, and soil nutrients cycling and reserving, less attention has been paid to changes about Bryophytes' niches and covers with climate warming on the QTP. Considering the global greenhouse gas emissions (SSP2-4.5), this paper delved the potential distribution, range shifts and key environmental variables of Bryophytes under current and future climate scenarios.
Geographically, the suitable habitats distributed in non-permafrost region of SE QTP and the area reached to 9.12 × 10 5 km 2 (covered 35.43% of total QTP). Without considering the biotic interactions, the different range shifts of Bryophytes species habitats mainly depended on habitat heterogeneity and environmental factors (Gao et al., 2016). In the larger scale, habitats with higher habitat heterogeneity were more conducive to the coexistence of species, which was shown in the spreading of gained Bryophytes to higher elevation regions, where usually more colder and wetter to be the considerable diversity (Sun et al., 2013;Zanatta et al., 2020). Additionally, the range shifts of Bryophytes were highly in uenced by external environments and interactions of multiply factors due to the speci c eco-physiological and biological features of Bryophytes (Mateo et al., 2013). Soil moisture, UV-B seasonality, temperature seasonality and precipitation of the coldest quarter were the four most predominant environmental factors by the Jackknife tests, and which were also compared with previous studies (Song et al., 2015;He et al., 2016;Tomiolo et al., 2018). Physiological of Bryophytes would change when exposed to UV-B seasonality, including changes of light quality, duration and intensity, in particular, a high UV-B radiation intensities at low temperatures, would be permanently damaged (Kallio and Valanne, 1975;Kershaw and Webber, 1986). Additionally, they preferred damp or humid habitats along with river (Levetin et al., 1996). So, it was common recognized that the peak photosynthetic activity of Bryophytes happened in early morning and late evening when the moisture conditions were most suitable (Bartels et al., 2018). In the meanwhile, the distribution and niches of Bryophytes were likewise coordinated by air temperature and its seasonal changes. Dilks and Proctor (1975) approved that there exist a relatively narrow temperatures extent for Bryophytes to obtain net photosynthesis. It seems that at high temperature, most Bryophytes on the QTP may suffer irreversible degradation or die (Weis et al., 1986;Löbel et al., 2018), but it could withstand much lower habitats in very cold climate for many years (Perera-Castro et al., 2020). Besides these three factors, precipitation de nes Bryophytes' growth and distribution within those boundaries. Thus, we can inferred that extreme climate events might be the main factors for range shifts Bryophytes on the QTP (Zhu et al., 2017;Yao et al., 2018;. Thus, there is a need for further studies to delve the larger variation in climate and more frequent climate extremes on Maxent model to improve the feedback of Bryophytes and niches to climate change. In addition to the climate change, the biomass and diversity of shrub and herbaceous vegetation (Jägerbrand et al 2012., Chen et al., 2017Fergus et al., 2017), freeze-thaw cycles for frozen soil (Porada et al., 2016a;Higgins et al., 2018) and excessive grazing (Olden et al., 2016), can also affect the distribution of Bryophytes. Thus, future work should synthesis other crucial factors into Maxent modeling for Bryophytes.

Conclusions
In this study, changes in potential geographical distribution of Bryophytes under current and CMIP6 future projections on the QTP were projected using Maxent model based on species occurrences and relative environmental variables. The in uences of environmental heterogeneity to habitats suitability were also analyzed. Our results indicated the occupied habitats of Bryophytes would increase slightly and move towards to the central and western regions with an overall surface air warming and moistening, solar dimming. The ndings in this study can be used to clearly understand the distribution and range shifts of Bryophytes, and provide a basis for habitats and their biota conversation of Bryophytes, even conductive for monitoring climate and ecosystem change on the degradation regions of permafrost and glacier in the central and western QTP.  Box plot showed the elevation change in four future time periods. '+'and '-'represent the gained and lost habitats of Bryophytes relative to the former time period, respectively. The lower boundary of the box indicates the 25%, a line within the box marks the 50% (median), and the upper boundary of the box indicates the 75%.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryMaterial.docx