Characteristics of temperature evolution from 1960 to 2015 in the Three Rivers’ Headstream Region, Qinghai, China

The spatiotemporal characteristics analysis

The temporal characteristics

As seen in Fig. 2a, the climate tendency rate of the annual mean temperature series is 0.337 °C per decade, and its correlation coefficient is 0.7674 (> r₀.₀₁ = 0.3357). In conjunction with the result of the M–K trend test, this finding demonstrates that the annual mean temperature has significantly increased in the past 56 years. In addition, the climate tendency rates of the annual mean temperature in the LARHR, YERHR and YARHR are 0.352 °C per decade, 0.34 °C per decade and 0.319 °C per decade, respectively (except for the climate tendency rate of the annual temperature in the YARHR, the correlation coefficients of the annual temperature series exceeded the significance level of 0.01). These rates exceed the mean rising rate of annual mean temperature in China (0.21–0.25 °C per decade) and that of the global annual mean temperature (0.07℃ per decade) in the past 56 years. The climate tendency rates of the annual mean temperature in the LARHR and YERHR were higher than those in Northwest China (0.32 °C/10a)^1,27, and the climate tendency rate of the annual mean temperature in the YARHR was similar to that in Northwest China. The higher annual climate tendency rate of the THRHR is related to the increase in the lowest night temperature in the study area²⁸ and the large amount of solar radiation in the lower altitude area of the THRHR¹³; furthermore, it may also be related to the decrease in total cloud cover and the increase in snow cover in the study area²⁹. In addition, as seen in Fig. 2a,b, the annual mean temperature of the THRHR has increased significantly since the late1990s.

As seen in Fig. 3, the annual and seasonal mean temperatures in the THRHR and its three subregions have been increasing in fluctuations since the 1960s, and the fluctuations in winter and spring are greater than those of summer and autumn. Due to the higher climate tendency rate of the summer, autumn and winter in LARHR than in the other two subregions, the orders of the climate tendency rate of seasonal mean temperature in the THRHR are completely consistent with that of LARHR (autumn, summer, winter, and spring mean temperatures), which are different from that in Northwest China (winter, autumn, spring and summer mean temperatures). In the three subregions, the spring climate tendency rate is lowest, and the orders of the climate tendency rate of autumn and winter are higher than that of spring, while that of summer varies greatly, it is highest in the YARHR, and that of autumn in the LARHR and the YERHR are highest. The high temperature rise rate in autumn and winter in the THRHR is related to the increase in the lowest temperature at night, the climate high tendency rate of winter in three subregions corresponds with the positive phase of the Arctic Oscillation (AO), and namely, the high climate tendency rate of winter corresponds with the high value of the AO. In addition, the annual and seasonal mean temperatures in the LARHR are ≥ 0 °C, respectively; meanwhile, the increasing range of the seasonal mean temperature climate tendency rate is significantly higher than that in the YERHR and YARHR, respectively. The decadal mean temperatures, the decadal minimum and maximum temperatures and their extreme values in the YERHR are lower than those in the LARHR, while they are higher than those in the YARHR (the value of decadal mean temperature is < 0 °C), and the increasing range of the decadal mean temperature in the YERHR is the smallest (Table 1). In comparison, the mean temperature in the YARHR with a higher average altitude is relatively low, while the mean temperature in the LARHR with a lower altitude is relatively high.

As seen in Fig. 3, the change in annual mean temperature in the three subregions tends to be consistent with that of the winter mean temperature and summer mean temperature, especially that of the summer mean temperature. The annual and the summer mean temperatures all have been increasing since the mid-1960s. Their increasing rangeability was relatively low in the 1970s–1980s and that of the mid-late 1980s was significant, especially since the mid-1990s. The spring mean temperatures have been increasing since the 1960s; they were relatively low in the 1970s–1980s (the decreasing rangeability of spring mean temperature in the YERHR was the highest), and they have been increasing obviously since the 1980s (especially since the mid-1990s). The autumn mean temperature has been increasing since the 1960s; it changed smoothly in the mid-1970s to 1980s and began significantly increasing in the 1980s (especially since the 1990s). The winter mean temperatures have been increasing since the early 1960s to early 1970s; they decreased from the mid-1970s to late-1980s, and they have been increasing significantly since the 1990s (especially since the mid-1990s). Furthermore, there was a slight cooling trend in approximately 2010. The annual and seasonal mean temperatures during 2010–015 were highest, while their increasing rangeability was smaller than those in the 1990s. Moreover, the occurrence of the weak downward trend of the summer mean temperature in the study area may be related to the local topographic fluctuations at high altitudes, which directly affect the local radiation balance and in turn affect the temperature change²⁸.

As seen in Table 2, the climate tendency rates of the annual mean temperature series in the three subregions are all lower than that of the Qinghai Lake basin from 1961 to 2018. Except that the climate tendency rates of the annual mean temperature series of the YERHR are roughly the same, those of the THRHR, the LARHR and the YARHR are higher than those in the north-western region, the Qinghai-Tibet Plateau and the southwestern region. Therefore, although the temperature change rate of the THRHR is lower than that of Qinghai Lake basin due to the various time periods, it is still a relatively warming area compared with the surrounding areas.

The spatial characteristics

The spatial distribution of the annual and seasonal mean temperature changes in the THRHR based on ANUSPLIN from 1960 to 2015 is shown in Fig. 4. This interpolation method comprehensively considers the role of concomitant covariates (such as altitude, latitude and longitude, etc.) in temperature change, and it is suitable for the spatial interpolation of meteorology series at high latitudes^30,31,32. As seen in Fig. 4, the spatial distribution of the annual mean temperature changes in the THRHR tends to be consistent with that of the winter and autumn mean temperatures, especially the former. (Their climate tendency rates are higher than those of annual, summer and spring mean temperatures).

Affected by the statistical samples and the topography³³, only the multiple correlation test between the climate tendency rate of autumn mean temperature and altitude exceeds the significance at α = 0.05 among the annual and seasonal mean climate tendency rates from 13 meteorological stations. In other words, the climate tendency rates in the three subregions of the THRHR do not increase consistently with the increase in altitude³⁴. As a whole, the climate tendency rates in the three subregions increase with increasing altitude from south to north and from southeast to northwest, and such a situation is much more obvious in the winter. This result aligns with the finding of Tandong Yao et al., who found that the climate warming rangeability in the Qinghai-Tibetan Plateau generally increased with altitude⁴. In terms of the three subregions of the THRHR, except for the climate tendency rate of the winter mean temperature, the annual mean temperature, spring mean temperature, summer mean temperature and autumn mean temperature in the LARHR are higher than those in the YERHR and YARHR, and the spring and autumn climate tendency rates of the YERHR are slightly higher than those in the YARHR; furthermore, their summer climate tendency rates are close, but their winter climate tendency rates are much higher than that in the YARHR.

The multiple linear regression equation is used to fit the relationship between the autumn and winter tendency rates (their climate tendency rates are relatively higher than other seasons) and the longitude, latitude, and altitude. The results show that the confidence test of the two fitted equations exceeds the significance level of 0.05, which again proves that in addition to atmospheric circulation, the topography has a certain impact on the spatial difference in temperature changes in the Qinghai-Tibetan Plateau³⁵. The factors affecting the climate tendency rate of autumn include altitude, longitude and latitude (the complex correlation coefficients between the climate tendency rates of autumn mean temperature and altitude and longitude passed the significance test of 0.01, while the multiple correlation coefficients between it and latitude failed the significance test), and those of winter include longitude, latitude and altitude (the complex correlation coefficients between the climate tendency rates of the winter mean temperature and longitude and latitude passed the significance test of 0.01, while the complex correlation coefficient between it and altitude only passed the significance test of 0.1). Furthermore, the correlations of the climate tendency rate of the autumn mean temperature with longitude, latitude and altitude are reversely in line with those of the climate tendency rate of winter, and the further research is needed to study the cause of this opposite trend.

The elevation effect in which the climate tendency rate of the autumn temperature increases with altitude in the study area shows that the melting of ice and snow in high-altitude areas may be affected by the increasing temperature, and the increasing temperature further aggravates the melting of ice and snow³⁶. Meanwhile, the surface albedo is effectively reduced, which results in the increasing LST, and then this will further speed up the melting of ice and snow, all of which will lead to the elevation effect on the mean rising rate of the LST. In winter, the surface albedo is high, and the LST is relatively low.

Source: Ecology - nature.com