This paper mainly was based on the average temperature, precipitation, humidity, and wind direction of Gangcha county from 1960 to 2013. By using wavelet analysis and Mann-Kendall (M-K) mutation analysis, specifically analyzed the climate change characteristics in the lake basin area of Gangcha county. The result showed that the climatic change in the lake basin area of Gangcha county is noticeable. The average temperature, average minimum temperature, average maximum temperature, and evaporation showed an increasing trend. But the evaporation in the study area was higher than precipitation. The average relative humidity showed a decreasing trend. And the sunshine and the average wind speed percentage showed a significant decreasing trend. Utilizing the Morelet wavelet, The time series of annual mean temperature, annual evaporation and annual sunshine percentage all have quasi3a and quasi4A periods, and the annual mean precipitation has quasi2–3A, quasi2–5A and quasi2–6A periods, which appear in 1996–2005, 1962–1978 and 1978–1996 respectively. The mean annual relative humidity has obvious quasi2–7A and quasi3A time series, which appear from 1960 to 1996, 1997 to 2005 and after 2008, respectively. The annual mean wind speed has quasi3–4A and quasi5A time series characteristics, which appear in 1964–1967, 1984–1995, after 2009 and 1971–1983, respectively. Through Mann-Kendall (M-K) mutation analysis, it is found that the mutations of evaporation and the average speed of wind are significant. The mutation of evaporation started in 2004, and the mutation of average started in 2003.

Global climate change has an increasingly significant and far-reaching impact on the world’s population, resources, and environment. As the core issue of global change research, climate change is an economic and political issue and has been a broad concern by the international communities. The Tibet Plateau, as the highest plateau in the world, is a “mirror” of the world’s climate change [

Morlet wavelet analysis and M-K test analysis are widely used for analyzing climate change. Wavelet analysis is a time-frequency analysis method with multi-resolution capability, which can be used in signal information processing and image processing [

Previous studies provided theoretical and methodological support for analyzing climate change characteristics, which showed different results in different study areas [

The above researchers studied the change characteristics of various climate factors in different time series through different methods. However, M-K test and Morlet wavelet analysis were not used to monitor the long-term change characteristics of all the climate factors in Qinghai Lake basin area. Lake basin area of Gangcha county is located in the northeast of Tibet Plateau, in the perennial westerly belt, is the intersection of the eastern Qinghai-Tibet Plateau monsoon region and climate region. Gangcha county’s unique geographical position makes lake basin area sensitive and vulnerable to climate change. Therefore, based on the annual and monthly data of average temperature, precipitation, relative humidity, sunshine percentage and wind speed from 1960 to 2013, this paper used Morlet wavelet analysis and M-K test to comprehensively analyze the characteristics and trends of climate change in this region. The study of the region is beneficial to understand the trend of global climate change, evaluate the regional ecological environment, and improve agricultural development [

The remainder of this paper is organized as follows. Section 2 is described the study area. Next, Section 3 should be showed research data and methods. Then, Sections 4, 5 and 6 are demonstrated and discussed the research results in detail, and finally, Section 7 presents the conclusions.

Gangcha county belongs to the Tibetan Autonomous Prefecture of Haibei in Qinghai Province, located in the northeast of Qinghai Province, southwest of Tibetan Autonomous Prefecture of Haibei, the north bank of Qinghai Lake. It is located between 99°20 ‘44 “−100°37’ 24” E and 36°58 ‘06 “−38°04’ 04” N. Gangcha county covers an area of 12,000 square kilometers, and the land type is mainly grassland. Gangcha county is a typical plateau continental climate, long sunshine time, a significant temperature difference between day and night; The average annual precipitation is 370.5 mm, and the annual evaporation is 1500.6∼1847.8 mm. Gangcha county is cool in summer and autumn and cold in winter. The yearly average temperature is only −0.6°C. In this paper, the south of lake basin area was selected as the research area (

This paper mainly adopted the annual average data and the monthly average data of relevant climate variables. The yearly average data were selected from the China Meteorological Data Service Centre (

Year | The average temperature (0.1°C) | Mean minimum temperature (0.1°C) | Mean maximum temperature (0.1°C) | Precipitation (0.1 mm) | Average wind speed (0.1 m/s) | Average relative humidity (%) | SAunshine percentage (%) | Small evaporation (mm) |
---|---|---|---|---|---|---|---|---|

1960 | −6 | −70 | 70 | 3225 | 34 | 54 | 68 | 1637.6 |

1961 | −9 | −73 | 63 | 3471 | 32 | 55 | 66 | 1575.9 |

1962 | −12 | −77 | 65 | 3119 | 34 | 52 | 70 | 1578.6 |

1963 | −7 | −72 | 70 | 4070 | 36 | 52 | 70 | 1528.8 |

1964 | −8 | −67 | 64 | 4073 | 34 | 58 | 67 | 1363.4 |

1965 | −7 | −68 | 66 | 3610 | 37 | 52 | 69 | 1366.2 |

1966 | −4 | −64 | 69 | 3286 | 38 | 52 | 66 | 1504.5 |

1967 | −15 | −71 | 53 | 4995 | 35 | 58 | 67 | 1214.8 |

1968 | −10 | −67 | 64 | 3454 | 39 | 58 | 69 | 1326.9 |

1969 | 0 | −62 | 76 | 3933 | 43 | 52 | 70 | 1484.4 |

1970 | −11 | −72 | 66 | 3693 | 40 | 51 | 69 | 1391.1 |

1971 | −3 | −63 | 71 | 4250 | 41 | 53 | 67 | 1525.6 |

1972 | −2 | −59 | 73 | 4083 | 38 | 55 | 67 | 1433 |

1973 | −2 | −64 | 73 | 2907 | 41 | 51 | 70 | 1542.8 |

1974 | −6 | −63 | 64 | 3930 | 39 | 57 | 66 | 1444.5 |

1975 | −8 | −65 | 61 | 4428 | 36 | 56 | 66 | 1448.2 |

1976 | −11 | −68 | 62 | 3578 | 32 | 54 | 70 | 1493.9 |

1977 | −9 | −66 | 63 | 3293 | 37 | 54 | 67 | 1538.5 |

1978 | −5 | −62 | 68 | 3033 | 39 | 54 | 70 | 1533.4 |

1979 | −1 | −62 | 73 | 3754 | 40 | 49 | 71 | 1743.5 |

1980 | −3 | −65 | 72 | 3257 | 39 | 50 | 71 | 1634.8 |

1981 | −1 | −60 | 72 | 4072 | 37 | 51 | 68 | 1549.6 |

1982 | −7 | −63 | 62 | 3659 | 36 | 56 | 69 | 1440.3 |

1983 | −12 | −68 | 58 | 3848 | 37 | 53 | 69 | 1321.6 |

1984 | −6 | −61 | 65 | 3507 | 39 | 52 | 69 | 1399.8 |

1985 | −5 | −59 | 65 | 4749 | 39 | 52 | 68 | 1403.5 |

1986 | −9 | −65 | 62 | 3848 | 35 | 53 | 71 | 1379.3 |

1987 | 2 | −55 | 75 | 4181 | 36 | 53 | 68 | 1460.6 |

1988 | 2 | −54 | 72 | 5158 | 35 | 56 | 64 | 1217.6 |

1989 | −5 | −57 | 62 | 4588 | 31 | 60 | 66 | 1136.9 |

1990 | 1 | −58 | 75 | 2601 | 34 | 52 | 70 | 1415.8 |

1991 | 0 | −61 | 74 | 3008 | 39 | 52 | 65 | 1470.7 |

1992 | −5 | −64 | 66 | 3089 | 36 | 53 | 65 | 1416.9 |

1993 | −5 | −61 | 66 | 3971 | 33 | 58 | 65 | 1363.8 |

1994 | 1 | −56 | 75 | 3711 | 34 | 57 | 67 | 1442.7 |

1995 | −6 | −63 | 68 | 3499 | 37 | 56 | 68 | 1502.2 |

1996 | −2 | −60 | 71 | 4097 | 35 | 56 | 68 | 1442.1 |

1997 | −2 | −59 | 70 | 3636 | 32 | 56 | 69 | 1410.8 |

1998 | 10 | −49 | 84 | 3801 | 33 | 55 | 68 | 1583.3 |

1999 | 5 | −52 | 79 | 4085 | 32 | 57 | 67 | 1443.3 |

2000 | −2 | −62 | 71 | 4196 | 31 | 56 | 69 | 1491.5 |

2001 | 4 | −57 | 78 | 2975 | 34 | 51 | 69 | 1536.6 |

2002 | 2 | −58 | 76 | 3722 | 33 | 56 | 70 | 1466.5 |

2003 | 7 | −53 | 80 | 3449 | 33 | 54 | 69 | 1481.6 |

2004 | 2 | −57 | 75 | 4279 | 33 | 52 | 70 | 1735.3 |

2005 | 6 | −51 | 76 | 4277 | 31 | 53 | 65 | 1755.3 |

2006 | 9 | −49 | 83 | 4155 | 31 | 54 | 66 | 1784.3 |

2007 | 6 | −50 | 80 | 4419 | 31 | 53 | 64 | 1866.7 |

2008 | 2 | −56 | 75 | 3958 | 30 | 52 | 65 | 1880.1 |

2009 | 7 | −48 | 79 | 4160 | 32 | 52 | 62 | 1775.2 |

2010 | 11 | −48 | 86 | 3968 | 33 | 49 | 67 | 1832.4 |

2011 | 5 | −50 | 77 | 4237 | 29 | 53 | 64 | 1705.9 |

2012 | −1 | −55 | 69 | 3712 | 28 | 53 | 58 | 1845.5 |

2013 | 11 | −51 | 77 | 3753 | 30 | 48 | 66 | 1763.4 |

2014 | 6 | −50 | 78 | 39 | 54 | 1655.6 |

Month | The average temperature (0.1°C) | Precipitation (0.1 mm) | Average relative humidity (1%) | Sunshine percentage (1%) | Average wind speed (0.1 m/s) |
---|---|---|---|---|---|

1 | −134 | 10.8 | 45 | 79.8 | 29.2 |

2 | −101 | 18.9 | 42.5 | 76.3 | 33.2 |

3 | −48 | 50.3 | 42.1 | 70.6 | 39.7 |

4 | 10 | 123.6 | 46.3 | 67.5 | 41.3 |

5 | 55 | 392.5 | 55.8 | 60.6 | 40.3 |

6 | 85 | 707.3 | 63.2 | 55.9 | 37 |

7 | 111 | 886.7 | 68 | 55.6 | 34 |

8 | 105 | 900.1 | 68.4 | 60.7 | 32.7 |

9 | 63 | 544.4 | 68.8 | 60.9 | 31 |

10 | 4 | 144.9 | 56.9 | 74.4 | 32.6 |

11 | −64 | 24.6 | 44.8 | 82 | 35.9 |

12 | −110 | 7.2 | 43.1 | 81.8 | 33.1 |

Moving average is the most basic method of trend fitting, which is equivalent to a low pass filter. For sequence X with a sample size of n, its moving average sequence is expressed as:

Wavelet analysis is a powerful statistical tool, which is widely used in many fields. The choice of function is critical when wavelet transform is a time series. Orthogonal wavelet functions are generally used for discrete wavelet transform, while non-orthogonal wavelet functions can be used for both discrete wavelet transform and continuous wavelet transform [_{0} is dimensionless frequency. When ω_{0 }= 6, the wavelet scale s is equal to the Fourier period (λ = 1.03 s) [

The continuous wavelet transform, the discrete-time series x_{n} (n = 1, 2, …, N) with isotime step δt is defined as the convolution of the wavelet function Ψ_{0} scale change as well as of the x_{n} under the transformation:

Wavelet full spectrum can show the unbiased and consistent estimation of the natural power spectrum of time series. The full wavelet spectrum can clearly identify the characteristics and intensity of periodic fluctuation of time series [

Although the time-frequency multi-resolution function of wavelet analysis can clearly reveal a variety of features hidden in time series, it is problematic to monitor the period of variation trend. When Jing Liu used Morlet wavelet analysis to analyze the period of measured annual runoff at various hydrological stations in northern Shaanxi province, It is found that the variation of annual runoff series has a certain influence on wavelet analysis period identification [

The Mann-Kendall test is a non-parametric statistical test method requiring sample data to follow a specific distribution and is not disturbed by a few outliers. It is suitable for the study of climate change in time series. A time-series x with a sample size of n constructs an order column r_{i}, r_{i} represents the accumulative sample count of x_{i }> x_{j} (1 ≤ j ≤ i). S_{k} is defined and calculated as follows:

Under the assumption of random independence of time series, define statistics:
_{k}) and Var(S_{k}) are the mean and variance of the accumulative quantity S_{K} respectively. UF_{i} follows the standard normal distribution, and UF_{1 }= 0 [

Put time series x in reverse order; it is arranged as x_{n}, x_{n−1}, …, x. Repeat the above steps while enabling UB_{k} = −UF_{k}, k=n, n−1, …, 1, UB = 0. Give the significance level, α = 0.05, u_{0.05} = 1.96, is plotted on the same plot as the UF_{k} and UB_{k} two statistical sequence curves. If the value of UF_{k} is greater than 0, the sequence is an upward trend. If the value of UFK is less than 0, the series shows a downward trend. When the two sequence curves exceed the critical straight line, it indicates a significant upward or downward trend. The range beyond the essential line is defined as the time region in which the mutation occurs. If the two curves intersect and the intersection is between the critical line, then the time corresponding to the intersection is when the mutation begins [

The average temperature in Gangcha county showed an obvious upward trend in the past 50 years (_{0.01 }= 0.345). The average temperature in Gangcha county increased at a rate of 0.3°C/10a in the past 55 years. The minimum value of the average temperature was −1.5°C in 1967, while the maximum value was 1.1°C in 2010 and 2013, and the extreme value was 2.6°C.

In recent decades, both the annual mean maximum temperature and annual mean minimum temperature in Gangcha county showed an increasing trend, and both passed the significance test with a confidence level of 99% (trend coefficient R = 0.631 > threshold r_{0.01 }= 0.345. The trend coefficient of annual mean minimum temperature R = 0.860 > threshold r_{0.01 }= 0.345). The warming rates of annual mean maximum temperature and the annual mean minimum temperature were 0.27°C/10a and 0.38°C/10a in the past 55 years respectively, and the extreme values were 3.3°C and 2.9°C respectively. It can be seen that although the warming rate of annual mean maximum temperature was lower than that of mean yearly minimum temperature, the warming range was higher than that of mean yearly minimum temperature (

The annual precipitation in Gangcha county showed a slight upward trend with an increasing rate of 5.6 mm/10a in the past 55 years, but this upward trend was not obvious and did not pass the significance test with a confidence level of 90% (trend coefficient R = 0.169 < threshold r_{0.1 }= 0.226). During the 1960 and 1970s, The fluctuation frequency of annual precipitation was relatively fast. In the 1980s, there was an obvious trend of rising fluctuation. In contrast, in 1990, the precipitation decreased rapidly, and after 1990, there was a trend of increasing slowly, and the annual average of precipitation was 381 mm, (

Evaporation in Gangcha county showed an overall upward trend, with an upward rate of 50 mm/10a in the past 55 years, which passed the significance test with a confidence level of 99% (trend coefficient R = 0.482 > threshold r_{0.01}= 0.345).

The average annual relative humidity in Gangcha County showed a slight downward trend from 1960 to 2014 (_{0.1 }= 0.224). The average relative humidity was 53.7% within the 55 years. Due to the influence of many factors, the interannual variation characteristics of the average relative humidity showed apparent fluctuation. Although there was an increasing trend of precipitation in Gangcha County, the increasing range was not as large as that of evaporation, which eventually had led to the decline of average relative humidity in Gangcha county.

The sunshine percentage in Gangcha county showed a decreasing trend year by year, with a decreasing rate of 0.64%/10a, in the past 55 years, (_{0.01 }= 0.345). The average sunshine percentage in Gangcha County was 67.5% within the 55 years.

The annual average wind speed in Gangcha county showed an obvious downward trend in the past 55 years (_{0.01 }= 0.345). The annual average wind speed in Gangcha county was 3.5 m/s. It was fluctuant for the trend of mean wind speed of every month within the 55 years. Starting from January, average wind speed increased gradually, to the peak in April, showed a clear trend of decrease, backed to the lowest in September, and then increased, to November, then decreased. There was a trend towards increasing first and then declining in the change mean wind speed of every month within the 55 years.

This section mainly used the Morlet wavelet power spectrum analysis method to study the periodic climate change rule in Gangcha county under the background of global warming.

There were significant quasi3a and quasi4a scale cycles in the time series of annual average temperature in Gangcha county (

Yuanfu Zhu in Gonghe county of Qinghai province, and analysis of the characteristics of climatic variation with the method of wavelet analysis, it was concluded that the annual mean temperature in Gonghe county were 47.0, 23.5, and 15.7 years change cycles, and had passed the test of significance of 99%, and This is different from the annual average temperature in Gangcha County, but the precipitation cycle is similar. Surong Guo pointed out in the analysis of temporal and spatial characteristics of climate change in Qinghai Province from 1960 to 2010 that the average annual temperature distribution in Qinghai province increased from south to north with a large difference between north and south, while the precipitation increased from northwest to southeast. Both Gonghe county and Gangcha county are lake basin areas of Qinghai Lake, and the time distribution of precipitation is roughly the same [

Through M-K test analysis, it can be seen from _{0.05} = ± 1.96) after 1988, indicating that this rising trend was very significant. This was consistent with previous analyses of interannual variations in average temperature. According to the intersection of the UF curve and UB curve, 1992 May was the abrupt change point of the warming trend in Gangcha county, but this abrupt change point was not within the confidence interval, and its credibility needs further discussion.

Evaporation in Gangcha county showed a fluctuated trend with time series, and the increasing trend was gradually intensified after 2004, and the warming trend greatly exceeded the critical line of the significance level of 0.05 (U_{0.05} = ± 1.96) after 2010, indicating that the increasing trend was significant. This is consistent with previous analyses of interannual variations in evaporation (

As can be seen from

The annual average wind speed in Gangcha county showed an obvious trend of increase and decrease, and the increasing trend was mainly before 1995, while the average wind speed showed a decreasing trend after 1995 (_{0.05} = ± 1.96), indicating that the decreasing trend was very significant. According to the intersection of UF and UB curves, we found that the decrease of mean wind speed in Gangcha county since the 21st century was a mutation phenomenon, and the mutation started in 2003.

Surong Guo analyzed the abrupt change of climate in Qinghai in recent 50 years and found that there was no obvious abrupt change trend of other climate factors except temperature. When Yufang Pei analyzed the abrupt changes of climate factors in Haidong city in 55 years, he found that the annual average temperature, evaporation, average wind speed, sunshine duration and relative humidity in Haidong City all had abrupt changes, indicating that the abrupt changes of climate factors in local areas were more significant than those in Qinghai Province. Therefore, it was normal for evaporation and wind speed in Gangcha county to have sudden changes.

Through a series of analysis of meteorological data in recent decades, the climate change characteristics in the lake basin area of Gangcha county was analyzed. The average temperature, average minimum temperature, average maximum temperature, and evaporation showed an increasing trend. The average temperature in Gangcha county increased at a rate of 0.3°C/10a. The minimum value of the average temperature was −1.5°C, while the maximum value was 1.1°C. Evaporation had an upward rate of 50 mm/10a. Precipitation increased slightly, with a variable rate of 5.6 mm/10a. The average relative humidity showed a decreasing trend year by year and the annual average relative humidity was 53.7%. Although the precipitation in the study area increased, the evaporation in the study area was higher than precipitation. The sunshine and the average wind speed percentage showed a significant decreasing trend, with the rates of −0.65%/10a and −0.13 m/s/10a, respectively. In general, the climate became drier and hotter, which would intensify the desertification process.

Through Morlet wavelet analysis, The time series of annual mean temperature, annual evaporation and annual sunshine percentage all have quasi3a and quasi4A periods, and the annual mean precipitation has quasi2–3A, quasi2–5A and quasi2–6A periods, which appear in 1996–2005, 1962–1978 and 1978–1996 respectively. The mean annual relative humidity has obvious quasi2–7A and quasi3A time series, which appear from 1960 to 1996, 1997 to 2005 and after 2008, respectively. The annual mean wind speed has quasi3–4A and quasi5A time series characteristics, which appear in 1964–1967, 1984–1995, after 2009 and 1971–1983, respectively.

Through M-K test analysis, it was found that evaporation and the mean wind speed had a significant mutation. The mutation of evaporation started in 2004, and the mutation of mean wind speed started in 2003. Although the average annual temperature changed significantly, the mutation point was outside the confidence interval, so it could not be judged that the change trend was significant. And the average annual precipitation, average relative humidity and sunshine percentage did not change significantly.

This work was supported by the Second Tibet Plateau Scientific Expedition and Research Program (STEP) under grant number 2019QZKK0804, and the National Natural Science Foundation of China “Study on the dynamic mechanism of grassland ecosystem response to climate change in Qinghai Plateau” under Grant number U20A2098.