Spatio-Temporal Changes of Snow Cover and Its Relationship With Meteorological Factors on the Qinghai-Tibet Plateau, China

The Qinghai-Tibet Plateau (TP) is one of the most sensitive areas to climate change, 13 and its ecological environment changes directly or indirectly reflect the global climate 14 change trend. The snow cover ratio (SCR) is an important indicator reflecting the 15 climate and environmental changes of the TP. The daily remote sensing data of snow 16 cover on the TP from 2003 to 2014 were used to study the spatio-temporal 17 distribution of snow cover on the TP. The results have shown that the average 18 snowmelt day on the TP starts on the 103rd day and ends on the 223rd day of a year, 19 and the snowmelt duration has a downward trend. Snow is mainly distributed in the 20 Nyainqentanglha Mountains, Karakoram Mountains and Himalayas. The SCR in 21 summer has a downward trend, while in autumn has a rising trend. This shows that 22 the difference in SCR during the year has enlarged, increasing the risk of snowmelt 23 floods. The SCR is highly correlated with temperature, but weakly correlated with 24 precipitation. Using the long-term remote sensing data of snow cover, the distribution 25 of glacier coverage on the TP can be extracted, in which glaciers on the TP account 26 for about 1%. This research provides an important reference for in-depth 27 understanding of the snow cover changes on the TP and their impact on the 28 environment.


Introduction
2019; J. Zhang et al., 2021). In climate change research, snow cover is a key climate 37 variable, and its seasonal variation is the most significant factor leading to changes in 38 surface albedo, which in turn causes the balance of earth and gas energy and regional and snow, which are an important water supply for rivers and affect the regional 65 surface water cycle, water resources distribution, and local production and life ( dataset was used to analyze the annual snow cover change, monthly snow cover 87 change, seasonal snow cover change, and correlation of snow cover with 88 meteorological factors. Furthermore, snow cover data was used to analyze the start 89 time and end time of snowmelt and extract glacier-covered areas. 90 2 Study area 91 The TP is called the third pole of the earth due to its cold natural environment. 92 Because it is the birthplace of many major rivers in Asia, such as the Yangtze River, 93 Yellow River, Yarlung Zangbo River, Lancang River, etc., it is also called the Asian 94 Water Tower. The geographical location of the TP is between 95 26°00′10″N-39°04′25″N, 73°03′37″E-104°07′59″E, with an area of 2.62 million km 2 .

96
The TP is surrounded by high mountains, with plateaus, basins and deep-cut canyons 97 in the middle ( . 102 The regional climate types of the TP are complex and diverse(J. Zhang et al., 2021).

103
The climate zones are mainly plateau temperate zone, plateau sub-frigid zone and 104 plateau frigid zone. Vegetation types change significantly with elevation and climatic 105 zones, and there are significant horizontal and vertical zoning characteristics in spatial 106 distribution, mainly including alpine shrubs, alpine grasslands and meadows. The TP 107 has long and cold winters, short warm summers, and large temperature difference 108 between day and night. The annual average temperature decreases from 20°C in the 109 southeast to below -6°C in the northwest. As the warm and humid air currents in the 110 southern ocean are blocked by multiple high mountains, the annual precipitation also 111 decreases from 2000 mm to less than 50 mm. The precipitation is most concentrated 112 in May to September and its spatial distribution is extremely uneven. The snowfall 113 starts in September, and the obvious snow cover period is from September to April of 114 the following year. The TP is also facing environmental problems such as glacier 115 melting, land desertification, and soil erosion. It is an area highly sensitive to global 116 climate change. the interpolated cloud removal algorithm is used for cloud removal processing to 126 obtain a daily snow-covered area cloudless product. The spatial resolution is 0.005 127 degrees and the temporal resolution is daily. The glacier catalog data of the 128 Qinghai-Tibet Plateau is extracted from the complete catalog dataset of global glacier 129 contours released by Global Land Ice Measurements from Space (GLIMS) 130 (http://www.glims.org/RGI/). 131

Snow cover days 132
Snow cover days (SCD) describes the number of times each pixel is covered by 133 snow in a certain period of time. The larger the number of snow cover days, the 134 longer the snow cover in the area and the more abundant the snow storage.
Where SCD is the snow cover days, N is the total duration time, is the judgment 137 value of daily snow cover for a certain pixel, 1 for snow cover and 0 for no snow 138 cover. 139

Snow cover ratio 140
The snow cover ratio (SCR) indicates the ratio of the snow cover time of a certain 141 pixel to the total time. The formula is calculated as follows: 142 Where is the duration time of snow cover, is the total duration time.  is plenty of water vapor and the temperature on the high mountains is low enough to 192 produce snow easily. Although the temperature in the northwest TP is low, the water 193 vapor is low, thus resulting in less snow. The number of SCD is similar to the spatial distribution of SCR. Generally, areas 198 with high SCR have more SCD. and low snow cover areas are relatively stable. Generally, in years with a high number 205 of snow cover days, snow cover mainly increases in the northeastern TP. 206 Fig. 7 illustrates the spatial distribution of the multi-year average SCR and SCD of 207 the TP. From Fig. 7 a, it can be seen that the SCR of 85% of the TP is less than 0.18. 208 The main areas with high SCR are concentrated in the southeast mountainous areas, 209 such as the Nyainqentanglha Mountain, Hengduan Mountain and Bayan Har 210 Mountain. As for the Karakoram Mountains in the west, the northern foothills of the 211 Himalayas in the southwest, and the southern foothills of the Qilian Mountains in the 212 northwest, these areas are mainly distributed with glaciers. From Fig. 7 b, it can be 213 seen that areas with high SCR generally have more SCD. The days of high SCD are 214 mainly concentrated in the Nyainqentanglha Mountains, Karakoram Mountains and 215 Himalayas. has the highest SCR, and August has the lowest SCR. These are also the two months 228 with the lowest and highest average temperature on the TP, indicating a strong 229 correlation between SCR and temperature. 230 SCR has an upward trend. In May, June, July and August, the SCR has a decreasing 236 trend, especially in June, where the decreasing trend is the fastest. This shows that the 237 snow cover rate of the TP is more differentiated during the year, and the SCR is 238 developing in a trend of steep rise and fall. The climatic environment has become 239 more differentiated. 240 241 Fig. 9 Changes in SCR of TP from January to December (a-l: Jan-Dec) 242 Judging from the changes of monthly average SCR (Fig. 10), February has the 243 highest SCR and August has the lowest SCR. The monthly average SCR is 0.141. The 244 SCR in January, February, March, November and December is higher than the 245 average level. The SCR in May and October is close to the average level, while in 246 June, July, and August , September it is below average. Snow melts mainly in May  247 and June, and accumulates in September and October. Nyainqentanglha Mountain is the most concentrated area of snow. In June, July, 256 August, and September, the snow melts in most areas, and the areas with snow cover 257 are mainly permanent glaciers. The SCR decreases in the fastest speed in April and 258 May, which is the main snowmelt period. The SCR increases in the fastest speed in 259 October, which is the main snow cover period. high in winter and spring, when the trends of SCR are not obvious and the state is 276 relatively stable. On the TP, the SCR is decreasing in summer, while increasing in 277 autumn. This shows that the seasonal SCR of the TP has become more differentiated, 278 with accelerated melting of snow from spring to summer, and a rapid increase in snow 279 from autumn to winter. This also indirectly indicates that the seasonal climate of the 280 TP is changing. 281 The spatial distribution of the average spring, summer, autumn and winter SCR 282 over the years is shown in Fig. 14 Himalayas, mainly because there are glaciers in these areas. The SCR in spring is 288 wide, but it is scattered, and the SCR in autumn is significantly larger than that in 289 summer. winter (d) on the TP 296

Relationship of SCR with precipitation and temperature 297
The annual precipitation distribution on the TP presents an obvious cascade 298 distribution, decreasing from southeast to northwest. Southern Tibet, the lower 299 reaches of the Yarlung Zangbo River, and the Hengduan Mountains have a lot of 300 precipitation, while the Qiangtang Plateau has sparse precipitation, with annual 301 precipitation less than 300 mm. Among them, 2008 has witnessed a wide area of large 302 precipitation, which is also the year with high SCR, while 2010 is a year with large 303 precipitation distributed in a small area, which is the year with low SCR. This 304 indicates that if there is more precipitation in a year, there will be more snow. In 305 addition, there is an increasing trend in areas with heavy rainfall in the southeast of 306 the TP. It is worth noting that in the Karakoram Mountains, although there is not 307 much precipitation, there is a large amount of snow cover, which is mainly due to the 308 widespread distribution of permanent glaciers in this area. The average annual temperature distribution on the TP is roughly in three levels 313 (Fig. 16). The area with the highest temperature is southern Tibet, the area with the 314 lowest temperature is the Qiangtang Plateau, and the temperature in other areas is 315 somewhere in between. Although the precipitation in southern Tibet is heavy, the 316 temperature is high, making the precipitation mostly in the form of rainfall. Therefore 317 there is little snow in this area. Although the temperature in the Qiangtang area is low, 318 the precipitation is scarce. Thus the snow cannot accumulate. Even if there is snow, it 319 will quickly evaporate. The Nyainqentanglha Mountain and the Hengduan Mountains 320 have low temperature and relatively high precipitation. The main form of precipitation 321 is snowfall, making these areas the most snow-covered areas. 322 Fig. 17 a illustrates the changes in SCR and annual precipitation from 2003 to 2014. 323 It can be seen from the figure that the annual precipitation trend is not obvious, while 324 the SCR fluctuates. From the analysis results of the correlation between precipitation 325 and SCR (Fig. 17 c), the correlation between precipitation and SCR is 0. Fig. 17 b  326 illustrates the change in SCR and annual average temperature from 2003 to 2014. It 327 can be seen from the figure that the SCR is low in the year with high temperature, and 328 is high in the year with low temperature. The SCR is negatively correlated with 329 temperature (Fig. 17 d), and the correlation is obvious, with a correlation coefficient 330 of 0.269. and correlations of SCR with precipitation (c) and temperature (d) 337

Distribution of glaciers on the TP 338
The Qinghai-Tibet Plateau and the marginal mountains are the most concentrated 339 areas of glaciers. Using Randolph Glacier Inventory (RGI) released by GLIMS, the 340 distribution of glacier cover on the Qinghai-Tibet Plateau is shown in Fig. 18 a. It can 341 be seen from the figure that glaciers are mainly distributed in the Nyainqentanglha 342 Mountains, Karakoram Mountains, Himalayas, Qilian Mountains and other regions. 343 The glacier coverage in the Karakoram area reaches 23.42%. Fig. 18 b is the 344 glacier-covered area obtained by statistics of snow cover. The area with a snow cover 345 rate greater than 0.9 is regarded as a glacier. Compared with Fig. 18 b, the glacier data 346 obtained through the snow cover data is more accurate, and the main glacier-covered 347 areas can be extracted. Glacier coverage on the Qinghai-Tibet Plateau is about 1.08%. 348 It provides a reference for the extraction of glacier cover. snow melt water accounts for 9.7% of the total runoff, the source area of the Yangtze 364 River accounts for 13.6%, and the upper reaches of the Heihe River accounts for 365 16.1%(Sobota et al., 2020). This shows that the snowmelt runoff mechanism in the TP 366 is obvious, and the impact of snow and glaciers on the hydrological cycle needs to be 367 considered. 368 Although the correlation between precipitation and SCR is 0 throughout the TP, it 369 cannot be denied that the correlation between precipitation and SCR is very high in 370 some areas. For example, the Nyainqentanglha Mountain area has a lot of 371 precipitation, which makes the SCR high, and the Qiangtang Plateau has less 372 precipitation, which makes the SCR low. In this study, the TP was taken as a whole to 373 analyze the relationship between precipitation and SCR, which will obscure some 374 local information. In future research, the TP can be divide into different regions for 375 study. (1) The average snowmelt on the TP begins on the 103rd day of the year, which is 385 early April, and the snowmelt ends on the 223rd day, which is early August. The 386 average snowmelt duration is 121 days, and the snowmelt time has a tendency to 387 shorten.

388
(2) The annual SCR fluctuates, and the main snow cover areas are the 389 Nyainqentanglha Mountains, Karakoram Mountains, and Himalayas. The SCR 390 decreases in May, June, July, and August, while it increases in April, indicating that 391 snow melting on the TP is accelerating. This is also reflected in the decrease in SCR 392 in summer and the increase in SCR in autumn, indicating that the difference in SCR 393 on the TP has enlarged during the year.

394
(3) The SCR of the TP is negatively correlated with temperature, but weakly 395 correlated with precipitation. However, there are differences in the correlation 396 between precipitation and SCR in different partial areas. Using remote sensing data of 397 long-term snow cover, the distribution of glacier cover on the TP can be extracted, 398 which is about 1%.   Location of study area and distribution of glaciers Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.   employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.  Spatial distribution of the multi-year average SCR (a) and SCD (b) on the TP Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Changes in SCR of TP from January to December (a-l: Jan-Dec) Figure 10 Changes in monthly average SCR on the TP Figure 11 Spatial distribution of monthly average SCR on TP from January to December (a-l: Jan-Dec) Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Spatial distribution of average SCR in spring (a), summer (b), autumn (c) and winter (d) on the TP Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.   Changes of SCR with precipitation (a) and temperature (b) from 2003 to 2014, and correlations of SCR with precipitation (c) and temperature (d) Figure 18 Glacier cover on the TP based on RGI (a) and SCR (b) Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.