Exploring temporal and spatial variability of ... - CyberLeninka

Journal of Hydrology: Regional Studies 9 (2017) 183–198

Contents lists available at ScienceDirect

Journal of Hydrology: Regional Studies journal homepage: www.elsevier.com/locate/ejrh

Exploring temporal and spatial variability of precipitation of Weizhou Island, South China Sea Shulin Deng a,b,c , Manchun Li a,b,c,∗ , Han Sun b , Yanming Chen a,c , Lean Qu a,c,d , Xianzhe Zhang a,b,c a b c d

Jiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Nanjing University, Nanjing, 210023, China Collaborative Innovation Center of South China Sea Studies, Nanjing University, Nanjing,210093, China Geographic And Oceangraphic Sciences, Nanjing University, Nanjing, 210023, China College of Territorial Resources and Tourism, Anhui Normal University, Wuhu, 241002, China

a r t i c l e

i n f o

Article history: Received 21 June 2016 Received in revised form 6 November 2016 Accepted 10 December 2016 Keywords: Rainfall trend Innovative method Mann-Kendall Weizhou island

a b s t r a c t Study region: Weizhou Island, Northern part of the South China Sea. Study focus: Research on precipitation variability is important for understanding the water cycle and to evaluate risk of flood and drought risks. We studied precipitation variability, including amounts and intensity, on a sea island (Weizhou Island) using a long time-series daily precipitation dataset. An innovative trend test (S¸en trend test) and M-K trend test were used to analyze trends of rainfall variability. The Concentration Index (CI1) and the precipitation concentration index (CI2) were used, respectively, to evaluate the intensity and seasonality of rainfall. The continuous wavelet transform (CWT) was used to analyze CI1 and CI2. NCEP/NCAR data were used to determine if large scale circulation has an influence on Weizhou Island rainfall variability. New hydrological insights: (1) Rainfall amounts had a non-homogeneous temporal distribution during periods of 1961–1990, 1981–2010 and 1961–2010 on Weizhou Island. (2) Large scale atmospheric circulation may be the major atmospheric driving force of precipitation changes. (3) Precipitation has a cyclical nature on Weizhou Island. (4) Precipitation pattern on Weizhou Island is also affected by oceanic climate. The results provide a scientific basis for water resource management on Weizhou Island. © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction Climate change and climate variability are critical topics. Atmospheric warming appears to have modified the water cycle, and precipitation plays a key role in this cycle (De Luis et al., 2011). Precipitation is a key element that directly affects water resource distribution which significantly influences water availability for agriculture, industry, and domestic uses. Heavy rain has great potential to trigger natural disasters, such as floods and landslides. These extreme natural events put substantial economic pressure on local and national governments (Caloiero, 2014; Li et al., 2014). Research on the variability of daily rainfall enables better understanding of the water cycle and enhanced evaluation flood and drought (Wang et al., 2011).

∗ Corresponding author at: Jiangsu Provincial Key Laboratory of Geographic Information Science and Technology, Nanjing University, Nanjing, 210023, China. E-mail address: [email protected] (M. Li). http://dx.doi.org/10.1016/j.ejrh.2016.12.079 2214-5818/© 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/ licenses/by-nc-nd/4.0/).

184

S. Deng et al. / Journal of Hydrology: Regional Studies 9 (2017) 183–198

Studies on precipitation variability have used various statistical methods. Increases or decreases in the number of rainy days have been detected in many areas (Sansom and Thompson, 2008; Norrant and Douguedroit, 2006; Gajic-Capka et al., 2015). Continental-scale trends in precipitation have been detected in Europe (Cortesi et al., 2012), central Asia (Xu et al., 2015), and also at national and regional scales (De Luis et al., 2011; Xu et al., 2003; Vincent et al., 2011; Guo et al., 2012; Alijani et al., 2008; Coscarelli and Caloiero, 2012). Precipitation variability on Weizhou Island, however, has not been studied. Extreme weather events, including droughts, floods and other secondary disasters, are likely to occur more frequently in the future (Coscarelli and Caloiero, 2012). Higher precipitation concentration, indicated by a high percentage of the total annual precipitation falling within a short period, has significant potential for triggering floods and causing droughts during subsequent lengthy dry intervals (Santos et al., 2010; Raziei et al., 2009). The amount and intensity of precipitation may increase slope instability and the risk of soil erosion. Soil erosion will have negative effects on growing conditions and agricultural practices, especially on Weizhou Island which has a fragile ecosystem. Therefore, it is important to analyze rainfall variability to provide a scientific basis for water resource management. The trend of rainfall amounts is an important aspect of rainfall variability. Changes of rainfall may alter groundwater recharge (Yu and Lin, 2015) and water availability (Hasan and Dunn, 2011). The S¸en trend test method developed by S¸en (2012) improves the ability to detect the trend degree (abrupt or gradual). Kisi and Ay (2014) used the S¸en trend test to analyze water parameters. They noted that the S¸en trend test had several advantages compared with the M-K trend test, and could be successfully used for trend analysis of low, medium and high values of water parameters. In addition, this method can also be used to analyze the trend of monthly pan evaporations. Kisi (2015) reported that the low, medium and peak pan evaporation values at some stations presented increasing and decreasing trends based on the S¸en trend test, but no trend was detected at the same stations using the M-K test. This innovative method was also applied by Ay and Kisi (2015) to analyze monthly total precipitation data in Turkey. Therefore, the S¸en trend test was used to analyze rainfall trends in this study. Rainfall amount is one important aspect of rainfall variability. Another important aspect of rainfall variability that merits consideration is intensity. The Concentration Index (CI1), established by Martin-Vide (2004), is used to analyze the variability of daily rainfall, and has been widely used to analyze precipitation concentration. Martin-Vide (2004) used CI1 to detect the spatial distribution of daily precipitation concentration in Spain. Peninsular Spain was divided into two regions using CI1 values: an eastern area of high concentration (where 25% of the rainiest days represent 70% or more of the annual total), and the rest of the country, which had more evenly distributed daily rainfall amounts. Higher precipitation CI1 values mainly occurred in Southern Xinjiang, whereas lower CI1 values were mostly found in Northern Xinjiang (Li et al., 2011). CI1 values are noticeably higher in regions where both annual total precipitation and number of rainy days are low. CI1 showed an inhomogeneous temporal distribution of daily rainfall in Southwest China (Shi et al., 2015). To quantify the heterogeneity of monthly precipitation in one year, Oliver (1980) proposed the precipitation concentration index (CI2). Shi et al. (2015) noted that Southwest China had a significant seasonality of rainfall distribution. CI2 values, derived from a daily time-series dataset of 1916–2006, in southern Italy ranged from a minimum value of 13.4 to a maximum value of 20.5, and illustrated the seasonality of the pluviometric distribution (Coscarelli and Caloiero, 2012). CI1 and CI2 are two descriptors of rainfall intensity, and both were used in this study. The trend in rainfall intensity is another important aspect of rainfall variability. The Mann–Kendall (M-K) trend method (Shi et al., 2015; Cortesi et al., 2012; Wang et al., 2013) has been widely used to study precipitation trends, CI1 tend, or CI2 trend (either increasing or decreasing), was also used. The goals of this study were to (1) evaluate trends of rainfall amounts at monthly, seasonal, and annual temporal scales; (2) study temporal patterns of CI1 and CI2 on Weizhou Island, (3) analyze the relationship between the variations of precipitation and large scale circulation patterns, (4) compare the rainfall variability of Weizhou Island with an adjacent mainland location (Malan, China). 2. Materials 2.1. Study area Weizhou Island (20◦ 54 –21◦ 10 E, 109◦ 00 –109◦ 15 N) is the youngest and largest volcanic sea island in the Northern part of the South China Sea (Fig. 1). It has an area of 24.7 km2 and a high elevation of 79 m. The annual average temperature is 23 ◦ C, and the annual average precipitation is 1350 mm. Due to tourism, much native vegetation has been replaced by buildings, and other structures. This has led to serious soil erosion by water runoff. Tourism has greatly increased the use of fresh water. The majority of fresh water on this island comes from precipitation. Affected alternately by the winter and summer monsoons, the temporal distribution of rainfall in Weizhou Island is inhomogeneous. Therefore, this island has a freshwater shortage most of the year around despite high annual rainfall. Understanding the rainfall pattern may provide a scientific basis to help local decision-makers better manage water resources. 2.2. Data A dataset of daily precipitation from 91 rain gauges in Guangxi province during the period 1961–2010 was obtained from the Meteorological Information Center of Guangxi. First, data quality and homogeneity were assessed. The purpose of data quality control was to identify errors in datasets of daily precipitation that might interfere with correct assessment of


185

Fig. 1. Location of the study area and spatial distribution of rain gauge.

the extremes (Vincent et al., 2011). The homogenization method reported in (Brunetti et al., 2012) was applied to analyze the homogenization of data series, and a two-step approach (Simolo et al., 2009) was used to estimate missing data. The quality test results of the database are not shown here. Daily precipitation records at Weizhou Island and Malan County were selected in this study. The spatial distribution of rain gauges on Weizhou Island is illustrated in Fig. 1. The NCEP/NCAR data from 1961 to 2010 was downloaded from http://www.esrl.noaa.gov/psd/data/grid. This dataset was used to analyze atmospheric moisture and related transport features. 3. Methodology CI1 and CI2 were used, respectively, to evaluate the daily and monthly rainfall pattern. The S¸en trend test and the M-K trend test were also used to study the trend of rainfall amounts and CI1/CI2. 3.1. Concentration index (CI1) To determine the relative impact of different classes of daily precipitation and to evaluate the relative importance of the largest daily event compared with the total daily events, it is necessary to analyze accumulated percentages of precipitation Y contributed by the accumulated percentage of days X when it occurs. These percentages are related to positive exponential curves, termed as normalized rainfall curves (Martin-Vide, 2004; Wang et al., 2013). X and Y are cumulative frequencies of precipitation and rainy days, respectively (Coscarelli and Caloiero, 2012), and are defined as follows: j

Yj = 100 ×

i=1 N j=1

j

pi Xj = 100 ×

pj

ni

i=1 N

(1) nj

j=1

where N is the number of classes divided by the 1 mm rainfall breakpoint, and pi and ni are the amount of rainfall and the number of rainy days falling into the ith class, respectively. The function of Y and X is: Y = aX exp(bX) (2)where a, b are constants, that are computed by the least-squares method.

186


After all of these constants have been confirmed, the area A is determined by the exponential curve and coordinate axis between 0 and 100 as follows: A=

a

b

ebx (x −

1 ) b

100

(3) 0

The daily precipitation CI1 resembles the Gini coefficient, and is the fraction of S over the area of the lower triangle delimited by the equi-distribution line (Martin-Vide, 2004). CI1 = S/5000

(4)

where S can be expressed as S = 5000 − A A high precipitation CI1 value indicates that precipitation is more concentrated within a few rainy days during the year, and vice versa. 3.2. Precipitation concentration index (CI2) CI2 indicates the distribution of monthly rainfall, Oliver (1980) proposed a definition of this index as follows: 12

CI2 =

pi 2

i=1 12 2 × 100

(6)

pi

i=1

where pi is the monthly precipitation falling into ith month. Oliver (1980) suggested that CI2 values less than 10 represent a uniform precipitation distribution, values from 11 to 15 indicate a moderate rainfall concentration, and values from 16 to 20 indicate an irregular precipitation distribution. CI2 values above 20 represent a significantly irregular precipitation distribution. 3.3. M-K trend test The M-K trend test, which is highly recommended for general use by the World Meteorological Organization, has been used to detect the trend and significance of many rainfall-related indices. In the M-K trend test, the null hypothesis (H0 ) is that the time-series data {xi , i = 1, 2, ..., n} are independent and identically distributed random variables, and the alternative hypothesis (H1 ) is that a trend exists in the time-series data (Huang et al., 2014). The statistic parameter, S, is defined as: S=

n n−1

sgn(xj − xk )

(7)

k=1 j=k+1

where n is the length of the time-series, k = 1, 2, ..., n − 1, j = 2, 3, ..., n, and sgn(xj − xk )can be expressed as

sgn(xj − xk ) =

⎧ +1 if(xj − xk ) > 0 ⎪ ⎨ ⎪ ⎩

0 if(xj − xk ) = 0

(8)

-1 if(xj − xk ) < 0

when n > = 8, S can be treated as normally distributed with a mean value of zero. If there are unrelated groups, V(S) is formulated as V(S) =

n(n − 1)(2n + 5) 18

(9)

Standardized statistic Z can be defined as

Z=

⎧ S−1 ⎪ S>0 ⎪ ⎪ ⎪ ⎨ V(S) 0S = 0

⎪ ⎪ S+1 ⎪ S2.576, at the 5% significance


187

Fig. 2. The innovative trend method proposed by S¸en (2012).

Fig. 3. Rainfall amounts from 1961 to 2010 on Weizhou Island.

level if |Z| >1.96, and at the 10% significance level, if |Z| > 1.645 (Huang et al., 2014). In the M-K test, the Kendall slope, which indicates the magnitude of the monotonic change (Xu et al., 2003), is another useful index formulated as follows: x −x ␤ = Median( jj−i i ), ∀i < j, 1 > i > j > n ∀i < j, 1 > i > j > n(11)where the estimator ␤ is the median of all recorded pairs for the entire dataset.

188


Fig. 4. Trend of rainfall amounts on Weizhou Island from 1961 to 2010.

3.4. S¸en trend test Effective and efficient water resources management requires not only trend identification, but also determination of whether changes are abrupt or gradual. In addition, the magnitude of the changes and length of the period have considerably different implications for planning and the management of water resources. S¸en (2012) proposed an innovative trend test to meet these requirements. The basis of the S¸en (2012) trend test is that two identical time-series show a scatter of points along a 1:1 gradient (45◦ ) line, when they are plotted against each other in the Cartesian coordinate system. However, if two non-identical time-series (with different values) are plotted against each other, data points are scattered either above or below the 1:1 (45◦ ) line. These plots allow the identification of three different trend types: no trend, increasing trend, or decreasing trend. For example, a hydro-meteorological or hydro-climatic time-series data is divided into two equal number sub-series (sorted in ascending order). The first sub-series (Xi ) comprise the X-axis, and the second sub-series (Xj ) comprise the Y-axis. Then, they are plotted using the Cartesian coordinate system (Fig. 2). Data points located on (or near) the 1:1 (45◦ ) line indicates no trend (trendless


189

Fig. 5. Trend of rainfall amounts on Weizhou Island from1961 to 1990.

time-series). However, if data points are located below the 1:1 line, they are interpreted as having a decreasing trend, and conversely, if located above the line, an increasing trend.

3.5. Continuous wavelet transform Continuous wavelet transform (CWT) is a mathematical method used to detect stationary and transient changes in a time series. The xn is assumed to be a time-series with equal time spacing (dt). 0 () is a wavelet function depending on the dimensionless ‘time’ with zero mean and being localized in both frequency and time (Torrence and Compo, 1998). In this study, the Morlet wavelet was used and is defined as:

0

2 /2

() = −1/4 eiω0 e−

(12)

190


Fig. 6. Trend of rainfall amounts on Weizhou Island from 1981 to 2010.

where ω0 is the non-dimensional frequency (Torrence and Compo, 1998). The CWT of xn with a scaled and translated version is defined as: Wn (s) =

N−1 n

xn ∗

(n − n) ıt s

(13)

where the (*) indicates the complex conjugate, ıt is the sampling period, and s is the CWT scales. By varying the wavelet scales and translating along the localized time index n, one determines both the amplitude of any parameters versus the


191

Table 1 Trend of rainfall amounts on Weizhou Island during three time periods. Period

annual Spring Summer Autumn Winter January February March April May June July August September October November December

1961–2010

1961–1990

1981–2010

Low

Mid

Peak

Low

Mid

Peak

Low

Mid

Peak

Note: = positive trend, = negative trend, = no trend.

scale and how this amplitude varies with time. The subscript 0 on has been dropped to indicate that this normalized. In this study, CWT was used to study the periodical characteristics of CI1and CI2.

has also been

4. Results and discussion 4.1. Rainfall amount trend on Weizhou Island Annual and seasonal rainfall amounts were analyzed on Weizhou Island (Fig. 3). It has to be pointed out that spring season is from March to May, summer is from June to August, autumn is from September to November, and winter is from December to February (of the following year). The lowest of annual rainfall amounts is about 635 mm in 1962, and greatest of annual rainfall amount is 2400 mm in 2008. Rainfall amounts in spring and winter is relatively constant with only slight variation. Although, the temporal distribution of annual rainfall amount increases slightly, there are large fluctuations during the study period. The same temporal distribution appears during summer, which plays a key role in fluctuation of the annual rainfall amount. China has had rapid economic growth, since 1978 leading to large emissions of anthropogenic aerosols and greenhouse gases into the atmosphere. These gas emissions may alter the variability of regional precipitation (Sarojini et al., 2016). On the other hand, the East Asian monsoon weaken in the late 1970s and early 1980s (Zhang et al., 2008), and reduced the northern movement of the precipitation belt. The trends of rainfall amounts at monthly, seasonal and annual time scales were analyzed using the S¸en trend test during three periods (1961–2010, 1961–1990 and 1980–2010). The results are reported in Figs. 4–6, respectively. Fig. 4 shows rainfall trends on Weizhou Island from 1961 to 2010. There are no trends detected in annual and seasonal rainfall amounts. However, rainfall amounts considering the yearly and summer aggregation show an upward trend at higher precipitation values, but only a moderate downward trend at higher values in spring, autumn and winter. In addition, a descending trend appears in January, April, August, and October, while an ascending trend occurs during May, June, July, September, and November. Anomalies also present at highest (peak) values in May, October and December. Fig. 5 presents the trend of rainfall amounts on Weizhou Island from 1961 to 1990. There are increasing trends in annual, autumn and winter rainfall amounts. Rainfall amount in summer shows a moderate downward trend at lower precipitation values, but a moderate upward trend at higher values. For monthly precipitation, a rising trend is seen during January, February, and September, while there is a declining trend in October. The trend of rainfall amounts during 1981–2010 was analyzed (Fig. 6). It is clear that there is a decreasing trend of rainfall amounts in the summer and winter periods, but no trend in spring. A descending trend appears at lower values of annual and autumn rainfall amounts, but an ascending upward trend occurs at higher values of annual and summer rainfall amounts. For monthly precipitation, a rising trend appears in September, while there are declining trends in January, February, April, June, July, and August. Rainfall amount trends at low, mid and peak values on Weizhou Island were also analyzed during periods of 1961–2010, 1961–1990, and 1981–2010 (Table 1). Within the same period, rainfall amounts at different temporal scales present quiet different trends at low, mid and peak values. Rainfall amounts at the same temporal scale also show different trends at low, mid and peak values during the three different periods. The inconsistency of trends at low and peak values may have led to

192


Fig. 7. Spatial distribution of the whole layer atmospheric moisture flux anomalies in East Asia between 1961 and 1990 and 1980–2010 (unit kg m−1 s−1 ). The red rectangle indicates the approximate location of Weizhou Island. (a) summer, (b) autumn.

the anomaly value of rainfall amounts during the three study periods. The trends at mid values seem to be highly affected by the length of the study period. To fully explore the disparity of rainfall amounts during the three studied periods, the trends in these periods are compared (Fig. 4–6). Trends of annual rainfall amounts are different during the three study periods. While there is no trend in annual rainfall amount during 1961–2010, the annual rainfall amount shows a prominent upward trend from 1961 to 1990 and a downward trend from 1981 to 2010. In summer, winter, and February, trends of rainfall amounts coincided with that of annual trend. A significant increasing trend appears in autumn rainfall amounts during the 1961–1990 and 1980–2010 periods, but only a slight trend is detected in 1961–2010. The same phenomenon appears in September. Besides, rainfall


193

Fig. 8. Wavelet analysis of annual CI1 (1961–2010) on Weizhou Island.

amounts in April and August present a descending trend in both 1961–2010 and 1980–2010, but no trend is detected in 1961–1990. Rainfall amounts in June and July present totally different trends during the three study periods. There is a rising trend in 1961–2010, a declining trend in 1980–2010, and no trend in 1961–1990. Anthropogenic forces accelerated during 1980–2010, introducing much more aerosols and greenhouse gas into the atmosphere. This process can increase the amount of precipitation. However, rainfall amounts in summer and autumn (playing major roles in the fluctuation of annual rainfall amount) has a rising trend during 1961–1990, but a downward trend in 1980–2010. This tendency mainly occurs in the summer and the autumn months. These results indicate that gas produced by human activity may have a minor influence on rainfall variability on Weizhou Island. To analyze the reason why rainfall amounts in summer and autumn present a different trend between 1961 and 1990 and 1980–2010, summer and autumn moisture fluxes were also analyzed using the NCEP/NCAR reanalysis dataset. Spatial distribution of the summer and autumn whole layer moisture flux anomalies in East Asia are shown in Fig. 7. It is clearly that there is a tendency from north to south in both summer and autumn moisture flux anomalies in northeast-southwest China, indicating a downward trend of northward moisture fluxes. A similar large scale circulation tendency over East Asia also was reported by Zhang et al. (2010), Wang et al. (2011) and Zhang et al. (2008). These results showed that the East Asia monsoons had weakened and northern movement of the precipitation was reduced. Weizhou Island is affected by a negative difference of moisture flux in both summer and autumn, so moisture flux decreases in summer and autumn between 1961 and 1990 and 1980–2010. The change of moisture flux budgets is consistent with the trend of rainfall amounts in both summer and autumn, indicating that large scale atmospheric circulation may be the main driving force for the change of precipitation on Weizhou Island. 4.2. The periodical characteristics and trend of CI1/CI2 on weizhou island The continuous wavelet transform was used to explore periodical variability of the CI1 and CI2 time-series. The periodical characteristics of precipitation CI1/CI2 time series can be illustrated as follows: a 2-year significant period implies the pattern of CI1/CI2 will appear every 2 years, a 3-year significant period implies the pattern of CI1/CI2 will appear every 3 years, and so on(Li et al., 2011). CI1, CI2 and the percentage of precipitation contributed by 25% of the rainiest days for Weizhou Island and Malan County were evaluated (Table 2). CI1, CI2, and the percentage of annual precipitation contributed by 25% of the rainiest days are quite high on both Weizhou Island and Malan County during the study years. The periodical characteristic of CI1 on Weizhou Island was analyzed (Fig. 8). CI1 in Weizhou Island has a 2–4-year significance period, characterized by a strong activity, during the period of 1965–1985, and has a 2–6-year noticeable period during 1988–2010. Fig. 9 shows the periodical characteristics of CI2 on Weizhou Island. CI2 on Weizhou Island has a 2–8-year significance period during the period of 1961–2010. CI1 at different temporal scales, detected throuth the M-K and S¸en trend test, were compared (Table 3). It has to be pointed out that an increasing trend in lower value of CI1/CI2 means a rising potential to trigger flood or cause drought, and a downward trend in lower value indicates a decreasing trend of potential risk. On the other hand, a rising trend in higher value of CI1/CI2 indicates highly potential to trigger heavy flood or cause drought, and a downward trend in higher value means a decreasing trend of potential risk. These physical meanings of CI1/CI2 are highly related to rainfall amount. On Weizhou Island, the results of the M-K test applied to CI1 show no significant trend at annual, seasonal and monthly time scales. CI1, however, shows an upward or downward trend at low or peak value at most temporal scales detected by S¸en trend test. In addition, CI1 has an increasing trend in annual, summer, autumn, March and July, and a decreasing trend in January, April, May and December. The indicator of seasonality precipitation (CI2), used to quantify the relative rainfall patterns, was also explored in Weizhou Island (Fig. 10). CI2 values range from 11.8 to 32.4 on Weizhou Island, indicating an irregular seasonal distribution of precipitation. CI2 shows no trend at lower value, and a decreasing trend at peak value on this island, which is highly affected by extreme weather.

194


Table 2 CI1,CI2, and the Percentage for each year on Weizhou Island and Malan County. Years

1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Weizhou Island

Malan County

CI1

CI2

Percentage

CI1

CI2

Percentage

0.78 0.74 0.74 0.75 0.76 0.78 0.76 0.73 0.81 0.70 0.79 0.73 0.81 0.78 0.73 0.76 0.80 0.73 0.78 0.81 0.83 0.76 0.75 0.83 0.77 0.75 0.74 0.73 0.75 0.74 0.78 0.71 0.82 0.74 0.77 0.72 0.64 0.76 0.72 0.76 0.73 0.74 0.67 0.75 0.77 0.76 0.70 0.74 0.77 0.80

13.56 16.02 23.93 17.18 14.09 19.87 15.44 15.33 19.03 15.52 21.61 32.35 22.49 14.51 15.03 19.35 20.12 15.88 19.07 18.15 19.00 18.54 12.63 20.13 25.97 19.74 15.72 22.46 12.26 11.88 21.57 14.10 20.90 26.38 19.24 14.92 14.32 16.85 17.24 16.78 16.57 22.23 24.29 18.01 19.27 21.47 14.87 18.45 16.52 19.96

85.29 80.91 83.49 85.05 84.84 85.44 83.58 81.43 87.74 78.41 86.52 82.45 90.36 86.80 80.66 84.45 88.14 82.25 85.88 88.87 89.99 84.49 82.36 91.00 85.69 84.89 83.27 81.32 83.43 82.69 84.33 78.84 88.05 82.74 84.33 81.00 74.38 85.02 78.76 83.69 82.65 82.70 75.96 82.11 85.06 84.34 78.73 83.55 85.57 87.54

0.75 0.72 0.72 0.77 0.76 0.79 0.72 0.76 0.76 0.69 0.82 0.73 0.69 0.77 0.73 0.74 0.75 0.72 0.76 0.76 0.78 0.72 0.71 0.81 0.72 0.72 0.76 0.78 0.78 0.67 0.76 0.72 0.72 0.76 0.77 0.77 0.61 0.77 0.71 0.76 0.77 0.74 0.75 0.70 0.77 0.78 0.74 0.73 0.79 0.76

13.61 18.27 23.38 19.20 12.80 24.62 15.91 28.06 22.41 14.95 20.41 18.93 16.67 22.31 15.10 18.68 18.35 14.12 15.29 18.07 22.18 14.89 14.61 15.98 17.51 19.36 24.68 20.45 14.57 11.30 26.99 15.43 16.29 25.80 20.32 16.79 14.88 14.89 18.84 20.44 20.04 19.55 24.15 23.40 18.88 22.90 16.61 19.90 18.70 22.93

83.69 79.91 82.30 85.77 83.73 85.75 80.83 84.10 84.44 78.10 88.52 82.50 78.96 84.23 81.71 83.00 83.47 82.58 84.85 84.43 83.95 80.14 79.82 87.97 80.97 81.10 83.71 85.21 84.87 76.19 84.18 80.79 80.46 84.34 85.01 85.54 71.83 84.85 80.39 84.55 84.28 81.62 83.59 78.25 85.37 84.96 81.66 82.41 85.96 84.08

Note: The Percentage indicates percentage of precipitation contributed by 25% of the rainiest days.

4.3. The comparison of rainfall variability between weizhou island and the closest inland (Malan) To study the effects of oceanic and mainland climate on the pattern of precipitation on Weizhou Island, the trends of rainfall amount in Malan (which is the closest mainland location to Weizhou Island), were also examined (Fig. 11). In Malan, there is no obvious trend in annual, autumn and winter rainfall amounts, but an increasing trend appears in spring and summer. For monthly rainfall amount, a declining trend appears in January and April, and a rising trend is observed during June and July. However, no trends are seen in annual and seasonal rainfall amounts on Weizhou Island. CI trends in Malan and Weizhou Island were also compared (Table 3). In Malan, CI1 trends detected by M-K trend test and S¸en trend test are similar to CI1 trends on Weizhou Island. However, CI1 detected by S¸en trend test presents different trends at low or peak value at same temporal scales on Weizhou Island and Malan. CI2 ranges from 11.3 to 28.1 in Malan (Fig. 10). Precipitation also shows seasonal characteristics in this County. According to the S¸en trend test, CI2 in Malan has an increasing trend, but there is no trend on Weizhou Island. These results suggest that pattern of precipitation on Weizhou Island may be affected


195

Fig. 9. Wavelet analysis of CI2 (1961–2010) on Weizhou Island.

Table 3 Comparison of CI1 trend results detected by the M-K and S¸en trend test. Periods

CI1 trend (Weizhou Island) M-K

annual Spring Summer Autumn Winter January February March April May June July August September October November December

CI1 trend (Malan)

S¸en

M-K

Low

Peak

Note: = positive trend, = negative trend, = no trend. The M-K test results |Z| < 1.645 means no trend.

Fig. 10. CI2 trends on Weizhou Island and Malan from 1961 to 2010.

S¸en Low

Peak

196


Fig. 11. Annual, seasonal and monthly rainfall trend results for Malan from 1961 to 2010.

by oceanic climate and the pattern has some special characteristics compared to mainland precipitation, though it is quite similar to the mainland. The spatial distribution of CI1 values from south–north in Xinjiang, China, shows higher CI1 values in the Southern part and lower CI1 values in the Northern part. Rainfall trends in most parts of Xinjiang show no significant trend according to M-K trend test (Li et al., 2011). In Hunan province, higher CI1 are mainly detected in northwest parts, but the central parts are dominated by lower CI1 (Huang et al., 2014). Low CI1 was mainly found in the southwest of Jiangxi province, while the scattered areas in the northern Jiangxi province are characterized by high CI1 (Huang et al., 2013). The remaining parts of Jiangxi province are classified as areas dominated by medium values of CI1. In Haihe River basin, annual precipitation decreased at all stations. Only seven stations (located in western and southeastern parts) have significant trends at the 90% confidence level. Summer, July, August and November have a strong decreasing trend, while significant increasing trends are observed during May and December (Wang et al., 2011). These results illustrate that rainfall variability is highly spatial-related. The specific effects of oceanic climate on the pattern of precipitation on Weizhou Island require further research.


197

5. Conclusion Understanding daily rainfall distribution is important due to its influence on extreme climate events and also for the efficient management of water resources. This research presents a case study of temporal distribution of daily rainfall on Weizhou Island, located in the Northern part of the South China Sea. The results indicate that (1) rainfall amounts have different temporal distribution during 1961–1990, 1981–2010 and 1961–2010 on Weizhou Island. (2) Large scale atmospheric circulation may be the main driving force affecting the variation of rainfall trends on Weizhou Island, according to the coherence of the change of moisture flux budgets and the trends of rainfall amounts in both summer and autumn. (3) Different trends are detected by the S¸en trend test or the M-K trend test at some same temporal scales during three study periods. (4) Precipitation on Weizhou Island has cyclical characteristic. The study results can help in the management of water resources and allow better prediction, risk assessment, and management of natural disasters caused by unpredictable daily rainfall. Conflict of interest We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted. Acknowledgements This work is supported by the Major Program of National Social Science Foundation (Grant No. 14ZDA078-5). Sincere thanks are given for the comments and contributions of anonymous reviewers and members of the editorial team. Appendix A. Supplementary data Supplementary data associated with http://dx.doi.org/10.1016/j.ejrh.2016.12.079.

this

article

can

be

found,

in

the

online

version,

at

References S¸en, Z., 2012. Innovative trend analysis methodology. J. Hydrol. Eng. 17 (9), 1042–1046. Alijani, B., O’brien, J., Yarnal, B., 2008. Spatial analysis of precipitation intensity and concentration in Iran. Theor. Appl. Climatol. 94 (1–2), 107–124. Ay, M., Kisi, O., 2015. Investigation of trend analysis of monthly total precipitation by an innovative method. Theor. Appl. Climatol. 120 (3–4), 617–629. Brunetti, M., Caloiero, T., Coscarelli, R., Gullà, G., Nanni, T., Simolo, C., 2012. Precipitation variability and change in the Calabria region (Italy) from a high resolution daily dataset. Int. J. Climatol. 32 (1), 57–73. Caloiero, T., 2014. Analysis of daily rainfall concentration in New Zealand. Nat. Hazards 72 (2), 389–404. Cortesi, N., Gonzalez-Hidalgo, J.C., Brunetti, M., Martin-Vide, J., 2012. Daily precipitation concentration across europe 1971–2010. Nat. Hazard. Earth Sys. 12 (9), 2799–2810. Coscarelli, R., Caloiero, T., 2012. Analysis of daily and monthly rainfall concentration in Southern Italy (Calabria region). J. Hydrol. 416, 145–156. De Luis, M., Gonzalez-Hidalgo, J.C., Brunetti, M., Longares, L.A., 2011. Precipitation concentration changes in Spain 1946–2005. Nat. Hazard. Earth Sys. 11 (5), 1259–1265. Gajic-Capka, M., Cindric, K., Pasaric, Z., 2015. Trends in precipitation indices in Croatia, 1961–2010. Theor. Appl. Climatol. 121 (1-2), 167–177. Guo, J., Guo, S., Li, Y., Chen, H., Li, T., 2012. Spatial and temporal variation of extreme precipitation indices in the Yangtze River basin, China. Stoch. Env. Res. Risk A 27 (2), 459–475. Hasan, M.M., Dunn, P.K., 2011. Entropy, consistency in rainfall distribution and potential water resource availability in Australia. Hydrol. Process. 25 (16), 2613–2622. Huang, J., Sun, S., Zhang, J., 2013. Detection of trends in precipitation during 1960–2008 in Jiangxi province, southeast China. Theor. Appl. Climatol. 114 (1–2), 237–251. Huang, J., Sun, S., Xue, Y., Zhang, J., 2014. Spatial and temporal variability of precipitation indices during 1961–2010 in Hunan Province, central south China. Theor. Appl. Climatol. 118 (3), 581–595. Kisi, O., Ay, M., 2014. Comparison of Mann-Kendall and innovative trend method for water quality parameters of the Kizilirmak River Turkey. J. Hydrol. 513, 362–375. Kisi, O., 2015. An innovative method for trend analysis of monthly pan evaporations. J. Hydrol. 527, 1123–1129. Li, X., Jiang, F., Li, L., Wang, G., 2011. Spatial and temporal variability of precipitation concentration index, concentration degree and concentration period in Xinjiang, China. Int. J. Climatol. 31 (11), 1679–1693. Li, J., Tan, S., Chen, F., Feng, P., 2014. Quantitatively analyze the impact of land use/land cover change on annual runoff decrease. Nat. Hazards 74 (2), 1191–1207. Martin-Vide, J., 2004. Spatial distribution of a daily precipitation concentration index in peninsular Spain. Int. J. Climatol. 24 (8), 959–971. Norrant, C., Douguedroit, A., 2006. Monthly and daily precipitation trends in the Mediterranean (1950–2000). Theor. Appl. Climatol. 83 (1–4), 89–106. Oliver, J.E., 1980. Monthly precipitation distribution – a comparative index. Prof. Geogr. 32 (3), 300–309. Raziei, T., Saghafian, B., Paulo, A.A., Pereira, L.S., Bordi, I., 2009. Spatial patterns and temporal variability of drought in western Iran. Water Resour. Manag. 23 (3), 439–455. Sansom, J., Thompson, C.S., 2008. Spatial and temporal variation of rainfall over New Zealand. J. Geophys. Res. 113 (D6). Santos, J.F., Pulido-Calvo, I., Portela, M.M., 2010. Spatial and temporal variability of droughts in Portugal. Water. Resour. Res. 46 (W03503). Sarojini, B.B., Stott, P.A., Black, E., 2016. Detection and attribution of human influence on regional precipitation. Nat. Clim. Change 6 (7), 669–675. Shi, P., Wu, M., Qu, S., Jiang, P., Qiao, X., Chen, X., Zhou, M., Zhang, Z., 2015. Spatial distribution and temporal trends in precipitation concentration indices for the southwest China. Water Resour. Manag. 29 (11), 3941–3955. Simolo, C., Brunetti, M., Maugeri, M., Nanni, T., 2009. Improving estimation of missing values in daily precipitation series by a probability density function-preserving approach. Int. J. Climatol. 30, 1564–1576. Torrence, C., Compo, G.P., 1998. A practical guide to wavelet analysis. B. Am. Meteorol. Soc. 79, 61–78.

198


Vincent, L.A., Aguilar, E., Saindou, M., Hassane, A.F., Jumaux, G., Roy, D., Booneeady, P., Virasami, R., Randriamarolaza, L.Y.A., Faniriantsoa, F.R., Amelie, V., Seeward, H., Montfraix, B., 2011. Observed trends in indices of daily and extreme temperature and precipitation for the countries of the western Indian Ocean. J. Geophys. Res. 116, 1961–2008 (D10). Wang, W., Shao, Q., Peng, S., Zhang, Z., Xing, W., An, G., Yong, B., 2011. Spatial and temporal characteristics of changes in precipitation during 1957–2007 in the Haihe River basin China. Stoch. Env. Res. Risk. A 25 (7), 881–895. Wang, W., Xing, W., Yang, T., Shao, Q., Peng, S., Yu, Z., Yong, B., 2013. Characterizing the changing behaviours of precipitation concentration in the Yangtze River Basin, China. Hydrol. Process. 27 (24), 3375–3393. Xu, Z.X., Takeuchi, K., Ishidaira, H., 2003. Monotonic trend and step changes in Japanese precipitation. J. Hydrol. 279 (1–4), 144–150. Xu, L., Zhou, H., Du, L., Yao, H., Wang, H., 2015. Precipitation trends and variability from 1950 to 2000 in arid lands of Central Asia. J. Arid. Land. 7 (4), 514–526. Yu, H.-L., Lin, Y.-C., 2015. Analysis of space–time non-stationary patterns of rainfall–groundwater interactions by integrating empirical orthogonal function and cross wavelet transform methods. J. Hydrol. 525, 585–597. Zhang, Q., Xu, C.-Y., Zhang, Z., Chen, Y.D., Liu, C.-L., Lin, H., 2008. Spatial and temporal variability of precipitation maxima during 1960–2005 in the Yangtze River basin and possible association with large-scale circulation. J. Hydrol. 353 (3–4), 215–227. Zhang, Z., Tao, H., Zhang, Q., Zhang, J., Forher, N., Hoermann, G., 2010. Moisture budget variations in the Yangtze River Basin, China, and possible associations with large-scale circulation. Stoch. Env. Res. Risk A. 24 (5), 579–589.