Spatial and Temporal Variations of Extreme Climate Events in Xinjiang, China during 1961-2010 ()
Received 19 March 2016; accepted 15 August 2016; published 18 August 2016
1. Introduction
At present, no doubt exists about global warming. The fourth assessment report of the Intergovernmental Panel on Climate Change [1] identified a warming trend of 0.13˚C/10a over the past 50 years. The warming trend in China has been 0.22˚C/10a [2] , with significant increases in annual mean temperature and heating rate [3] . The increasing mean temperature not only affects the change in extreme temperature values, but can also leads to extreme climatic events, such as heat waves and rainstorms, which show trends of increasing frequency, strength, and intensity [4] .
Many studies have indicated that frequent extreme climatic events cause huge losses for society and the economy, as well as loss of human lives [5] . The most recent statistics reveal an increase of up to 10-fold in the economic loss arising from global climatic change and related extreme climatic events over the past 40 years. In China, weather disasters caused by extreme climatic events account for 70% of all natural hazards [6] . Hence, extreme climatic events are sources of widespread concern for governments and researchers [7] [8] . Studies of changes in these events are necessary not only to guarantee state security and economic development, but also to prevent disasters, reduce damage-related needs, and, most importantly, safeguard people’s safety and livelihoods.
Xinjiang, located far from the ocean in the center of Eurasia, in the border region of northwestern
China
, is impacted by the uplifted Tibetan Plateau, westerly wind circulation effects, and high mountain landforms. This region is thus an important route by which cold air invades
China
, and has a diverse climate. Northern Xinjiang has a variable continental arid and semi-arid climate, whereas southern Xinjiang has a warm continental arid climate; the ecology of these regions is fragile and has suffered greatly under the influence of climatic change [9] - [12] . In recent years, large studies have investigated the spatial and temporal distribution of and variation in extreme precipitation in Xinjiang [13] . These studies are of great importance for attaining a correct understanding of climatic variation in this region. However, most previous studies have considered single meteorological elements, such as temperature or precipitation in southern or northern Xinjiang, and discrepancies exist in the length of time represented by meteorological data and stations from which they were collected. Hence, this study was conducted using internationally recognized extreme climatic indices to examine long-term trends of change in extreme climatic events and their spatial distribution patterns, as well as problems existing in the whole of Xinjiang that have actual reference value for economic development and response to climatic disasters in the future.
2. Research Data and Methods
2.1. Research Data Source and Treatment
Daily precipitation and maximum, minimum, and mean temperature data from China’s Meteorological Administration for the period 1961-2010 in Xinjiang were used for this study. Data from 43 meteorological stations were reviewed to identify problems with missing or misdetected observations. We selected high-quality data from 35 stations (Table 1, Figure 1) as the study target. For missing and misdetected data from stations with less complete records, we used corrected data from Bai Lei et al. [14] .
2.2. Study Method
The standards used to define and calculate the extreme climatic indices used in this study were based on the World Meteorological Organization’s Commission for Climatology World Climate Research and Climatic Variation and Predictability programs’ expert team on climate change detection, monitoring, and indices [15] . This methodology has been applied to many countries and regions [16] , including Xinjiang. Using RClimDex software (http://cccma.seos.uvic.ca/ETCCDMI/), 19 extreme climatic indices (7 related to precipitation and 12 related to temperature) were developed and applied (Table 2).
All of the extreme climate indices reflect three aspects of temperature or precipitation events: strength, frequency, and temporal duration of climatic change. In extreme climate research, a certain percentile value is usually adopted as a threshold (with values exceeding it defined as extreme), and calculations are performed according to the nonparametric method of Bonsal et al. [17] , as follows: for a meteorological factor with N values, these values are sorted in ascending order (X1, X2, …, Xm, …, XN); the probability that a certain value is less than or equal to the rank of Xm is then P = (m − 0.31)/(N + 0.38), where P denotes percentile, m is the serial number of Xm, and N is the number of meteorological values.
In this study, percentile values were used to calculate relative and other indices. Specifically, temperature data were sorted in ascending order, and the 90th and 10th percentile values were regarded as thresholds for extreme temperature. When the highest temperature on a given day exceeded the 90th percentile value, an extreme high-temperature event was considered to have occurred on that day; when the highest temperature on a given day was less than the 10th percentile value, an extreme low-temperature event was considered to have occurred. Secondly, we analyzed extreme climatic events using linear tendency estimation, the Mann-Kendall method, and the kriging method of spatial analysis.
Table 1. Site meteorological information in the study area.
Figure 1. Distribution of meteorological stations in Xinjiang.
Table 2. Definition of extreme temperature and precipitation indices.
Notes: aAll the indices are calculated by RCLimDEX. Abbreviations are as follows: TX, daily maximum temperature; TN, daily minimum temperature; TG, daily mean temperature; RR, daily precipitation. A wet day is defined when RR ≥ 0.1 mm, and a dry day is defined when RR < 0.1 mm. Indices are included for completeness but are not analyzed further in this article.
3. Changes in Extreme Temperature Events in Xinjiang over the Past 50 Years
3.1. Interannual Variation in Extreme Cold Events
Consistent with global changes, the frequency of extreme low-temperature events in the Xinjiang area was found to have decreased while that of extreme high-temperature events increased [18] . Linear variation in extreme cold indices in Xinjiang over the past 50 years generally showed a decreasing curvilinear trend (Figure 2). The interannual variation tendency rates for the maximum numbers of continuous frost days (FDs), icy days (IDs), cold-nighttime days (TN10p), and cold-daytime days (TX10p) were −4.8 d/10a, −2.75 d/10a, −2.24 d/10a, and −0.88 d/10a, respectively (all α ≤ 0.01), and their correlation coefficients by year were −0.774, −0.475, −0.827, and −0.39, respectively (Table 3).
(Note: the straight line means linear fitting trend line, the smooth curve is moving average value in 5 years, attached mark is actual test value)
Figure 2. Regional annual anomalies series during 1961-2010 for indices of cold extremes.
Table 3. The extreme temperature mean value (˚C or d) and change range [˚C/
10a
)/or d/(
10a
)].
The monthly maximum temperature minimum value (TXn) and monthly extreme minimum temperature (TNn) showed increasing trends, with interannual variation tendency rates of 0.16˚C/
10a
and 0.505˚C/10a, respectively, although these trends were not significant (α > 0.05). The correlation coefficients of TXn and TNn by year were −0.133 and 0.042, indicating significant warming, especially on low-temperature days, the frequency of which has decreased over time.
3.2. Spatial Variation in Extreme Cold Events
Against the background of global warming, the extreme cold indices generally showed decreasing trends over the past 50 years, but changes in these indices were characterized by spatial variation (Figure 3). Change tendency rates of continuous FD ranged from −7.8 to −2.0 d/
10a
and displayed significant decreasing trends at about 91.4% of all stations (only 2.8% of stations showed increasing trends or no obvious trend). Variation tendency rates for IDs ranged from −4.0 to −1.5 d/
10a
, and were <0 at 80% of stations. Thus, the ID data showed significant decreasing trends in most regions, all of which were in northern Xinjiang.
Variation tendency rates for TN10p ranged from −3.7 to −0 d/
10a
and were <0 at 88.5% of stations; however, the decreasing trends were not significant. The rates for TX10p ranged from −1.3 to −0.5 d/
10a
and were < 0 at 48.5% of stations; these declining trends were significant. The variation tendency rates for TXn ranged from 0.04 to 2.5˚C/
10a
and exceeded 0 at 48.5% of stations; increasing trends were significant at 37.5% of stations. Variation tendency rates for TNn ranged from 0 to 2.5˚C/
10a
and showed obvious increasing trends at 94.2% of stations.
3.3. Interannual Variation in Extreme Warm Events
For the past 50 years, all extreme warm event indices, such as the numbers of summertime days (SU25), warm- nighttime days (TN90p), and warm-daytime days (TX90p), as well as monthly extreme maximum temperature (TXx) and the monthly minimum temperature maximal value (TNx), showed significant increasing trends in the Xinjiang area (Figure 4). The increasing rates for SU25, TN90p, and TX90p were 2.59, 3.38, and 1.59 d/10a, respectively (all α ≤ 0.05), with correlation coefficients by year of 0.48, 0.86, and 0.6, respectively (all with ≥99% confidence). The variation tendency rates for TXx and TNx were 0.23˚C/10a and 0.52˚C/10a, respectively, and the correlation coefficients by year were 0.36 and 0.64, respectively (all α ≤ 0.01).
The graph of changes in day-by-day temperature range (DTR; Figure 3 & Figure 4) shows a decreasing trend in Xinjiang over the past 50 years (rate, 0.21˚C/10a), with a correlation coefficient by year of −0.54 (α = 0.01). The trend of change in extreme annual temperature range in the entire Xinjiang area complies with that throughout China [19] .
3.4. Spatial Variation in Extreme Warm Events
Against the background of global warming, the extreme warm indices in the Xinjiang area generally showed increasing trends over the past 50 years, but these changes were characterized by spatial variation (Figure 5). Variation tendency rates for SU25 ranged from0 to 8.0 d/
10a
, and about 100% of stations showed significant increasing trends. The variation tendency rates for TN90p ranged from −0.2 to 8.0 d/
10a
, and 62.8% of stations showed significant increasing trends. The variation tendency rates for TX90p ranged from 0.2 to 3.2 d/
10a
and exceeded 0 at 91% of stations, with 80% of stations showing significant trends. The variation tendency rates for TXx ranged from −0.24 to 0.9˚C/
10a
and exceeded 0 at 77.1% of stations, with very significant increasing trends at 59.2% of stations. The variation tendency rates for TNx ranged from −1.86 to 3.0˚C/
10a
and exceeded 0 at 82.8% of stations. Trends of increasing TNx were very significant at 58.6% of stations.
Although the ranges of increase in variation tendency rates for TXn and TNn exceeded those for TXx and TNx, all stations showed decreasing trends for DTR.
(Note: the straight line means linear fitting trend line, the smooth curve is moving average value in 5 years, attached mark is actual test value)
Figure 4. Regional annual anomalies series during 1961-2010 for indices of warm extremes.
4. Changes in Extreme Precipitation Events in Xinjiang over the Past 50 Years
4.1. Interannual Variation Trends for Extreme Precipitation Indices
Precipitation intensity is one factor used to measure extreme precipitation, with greater intensity associated with greater possibility of disaster. Average precipitation intensity in the Xinjiang area showed a predominant trend of annual increases over the past 50 years, consistent with global data and those for China [20] . However, the range of change indicated by the extreme precipitation intensity indices was small relative to those for extreme temperature events. Total precipitation during the wet season (PRCPTOT) in each year from 1961 to 2010 appeared to show a significant increasing trend (rate,
8.23 mm
/
10a
; α = 0.05; Figure 6). The correlation coefficient of PRCPTOT by year was 0.43 (α = 0.01; Table 4). Average PRCPTOT over the years was
144.33 mm
(standard difference, 23.297). The difference between the two periods of 1961-1980 and after 1980 was very significant.
Note: the red dot means actual measure value, green line means moving average in 5 years, the straight line means simple regression trend line.
Figure 6. Inter-annual variation of precipitation extremes in Xinjiang during 1961-2010.
Table 4. The standard difference of extreme precipitation index and change tendency rate (mm/10a) and mean value (d or mm) in Xinjiang.
The simple precipitation intensity index (SDII) showed small increasing trend at a rate of
0.13 mm
/
10a
, but this trend was not significant (α > 0.05; Figure 6). The standard difference in the SDII was 0.549, the mean value was
4.62 mm
/d, and the correlation coefficient by year was 0.356, indicating that changes in the SDII were stable over the past 50 years (Table 4). All extreme precipitation indices, such as monthly maximum precipitation on 1 day (RX1day), monthly maximum precipitation in 5 days (RX5day), strong rainfall amount (R95), and number of consecutive wet days (CWDs), showed increasing trends; the rates of increase for the indices mentioned above were
0.47 mm
/
10a
,
0.43 mm
/
10a
, 9.32 d/
10a
, and 0.15 d/
10a
, respectively. Among these values, the increasing trends for R95 and CWDs were significant. The correlation coefficients by year for these indices were 0.25, 0.06, 0.35, and 0.41, respectively, with those of R95 and CWD reaching 95% confidence. The number of consecutive dry days (CDDs) showed a decreasing trend (rate, 2.87 d/
10a
), with a correlation coefficient by year of −0.29 (α = 0.05).
4.2. Spatial Distribution Pattern of Extreme Precipitation Indices
Interannual variation tendencies for most extreme precipitation indices, with the exception of the CDD index, showed significant increasing trends at most stations in the Xinjiang area over the past 50 years (Figure 7). PRCPTOT showed decreasing trends at only 2.8% of the 35 stations from which data were analyzed; most stations showed increasing trends (range, −30.68 to
19.78 mm
/
10a
). The rate of increase was greater in the Tianshan and northern Xinjiang areas. Only 17.1% of stations showed decreasing trends for SDII; 82.8% of stations showed increasing trends, and the rate of increase was greater at southern stations.
Increasing trends in the extreme precipitation indices of RX1day, RX5day, R95, and CWDs were found at 25 (71.4%), 28 (80%), 30 (85.7%), and 29 (82.8%) stations, respectively. More than 70% of stations showed increasing trends for PRCPTOT, SDII, RX1day, RX5day, R95, and CWDs, indicating that the frequency of extreme precipitation trends is generally increasing in the Xinjiang area. Analysis of CDD data from 11 (31.4%) stations in southern and northern Xinjiang revealed an increasing trend and demonstrated regional tendencies for increased numbers of disasters involving drought.
5. Conclusions
In this study, 12 extreme temperature indices and 5 extreme precipitation indices were used to examine temporal and spatial variation in climatic extremes in Xinjiang over the past 50 years. The main conclusions are described below.
1) Analysis of temporal changes revealed decreasing trends in extreme cold indices (i.e., obvious reductions in the number of severely cold days and extreme low-temperature events), consistent with global warming. Linear variation in extreme warm indices showed notable increasing trends. The trends for TX10p, TN10p, IDs, and FDs (−0.88, −2.24, −2.75, −4.8 d/10a, respectively) decreased, whereas those for TX90p, TN90p, and SU25 (1.59, 3.38, and 2.59 d/10a, respectively) increased.
2) Analysis of temporal changes in extreme precipitation indices (except CDDs), such as RX1day, RX5day, R95, and CWD, as well as PRCPTOT, showed consistently increasing trends (rates of 0.474 mm/10a, 0.434 mm/10a, 9.322 d/10a, 0.154 d/10a, and 8.232 mm/10a, respectively).
3) Differences in the spatial distributions of the indices were notable. The frequencies of extreme cold and extreme warm events decreased in southern Xinjiang. The spatial distribution of extreme precipitation also showed obvious regional differences, with the directionality of trends differing between mountainous and desert basin areas. The response to global warming has been more notable in northern than in southern Xinjiang. Global climatic change has altered the ecology of northern Xinjiang and increased the frequency of extreme climatic disasters.
4) Changes in extreme cold and warm indices, as well as those in nighttime and daytime indices, showed notable asymmetry: the warming ranges of the cold and nighttime indices exceeded those of the warm and daytime indices, respectively.
Acknowledgements
This study is jointly financed by the National Natural Science Fund Project (U1203282) and the National Natural Science Foundation Project of China (41001020), and the Xinjiang Uygur Autonomous Region Key Laboratory of “Xinjiang laboratory of Lake Environoments and Resources in Arid Zone” (XJDX0909-2012- 12), the Shihezi University team innovation project (2014ZRKXJQ08). The authors gratefully acknowledge funding for this research and would like to express their sincere thanks to Zhang Yan Wei for the help with data.
NOTES
*Corresponding author.