SpatiotemporalChangesandFrequencyAnalysisof …downloads.hindawi.com/journals/amete/2019/6324878.pdf · 2019. 8. 13. · O5qD O5λD O5PD O55D XDDD XDOD (b) j(– Fwfq ff5] ff(] ffX]

Research ArticleSpatiotemporal Changes and Frequency Analysis ofMultiday Extreme Precipitation in the Huai RiverBasin during 1960 to 2014

Yixing Yin 1 Xin Pan2 Xiuqin Yang1 Xiaojun Wang3 Guojie Wang 4

and Shanlei Sun1

1School of Hydrology and Water Resources Nanjing University of Information Science and Technology Nanjing 210044 China2Quzhou Meteorological Administration Quzhou 324000 China3State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering Nanjing Hydraulic Research InstituteNanjing 210029 China4School of Geographical Sciences Nanjing University of Information Science and Technology Nanjing 210044 China

Correspondence should be addressed to Yixing Yin yyxrosby126com

Received 14 January 2019 Revised 23 April 2019 Accepted 9 July 2019 Published 14 August 2019

Academic Editor Harry D Kambezidis

Copyright copy 2019 Yixing Yin et al +is is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Floods and droughts are more closely related to the extreme precipitation over longer periods of time +e spatial and temporalchanges and frequency analysis of 5-day and 10-day extreme precipitations (PX5D and PX10D) in the Huai River basin (HRB) areinvestigated by means of correlation analysis trend and abrupt change analysis EOF analysis and hydrological frequency analysisbased on the daily precipitation data from 1960 to 2014 +e results indicate (1) PX5D and PX10D indices have a weak upwardtrend in HRB and the weak upward trend may be due to the significant downward trend in the 21st century (2) the multiday (5-day and 10-day) extreme precipitation is closely associated with flooddrought disasters in the HRB and (3) for stations ofnonstationary changes with significant upward trend after the abrupt change if the whole extreme precipitation series are used forfrequency analysis the risk of future floods will be underestimated and this effect is more pronounced for longer return periods

1 Introduction

According to the IPCCrsquos fifth assessment report from 1880to 2012 the global average surface temperature has risen by085degC and global warming has become an undisputed factDue to global warming the water cycle has intensifiedleading to frequent weather and climatic events such as heatwaves drought and heavy precipitation Extreme weatherand climate events have become a focus for many scholars inrecent years because of their suddenness and great harm tohuman lives and property Both the intensity and occurrencefrequency of extreme precipitation have been changing withupwarddownward trends in different regions of the world[1ndash5] In recent years many researchers have also studied theextreme precipitation in China and have achieved a lot ofresults Zhai et al [6] tested on the trend of extreme

precipitation in China and found that there was no obviouschange in annual precipitation in China as a whole and theprecipitation intensity showed an upward trend but thetrend of extreme precipitation events showed obvious spatialdifferentiation

+e hydrological elements have changed a lot in re-sponse to the global climate change in the Huai River basin(HRB) [7ndash10] +e extreme precipitation events in this re-gion have also become one of the hot topics of scholars inhydrology and meteorology Xia et al [11] showed that mostof the maximum precipitation events concentrated in the1960ssim1970s most of which occurring in the flood seasonand the maximum daily precipitation increased in the HRBduring the period from 1960 to 2009 Lu et al [12] found thatthe precipitation events tended to be more extreme in1960ndash2008 in the HRB Yang and Cheng [13] found that the

HindawiAdvances in MeteorologyVolume 2019 Article ID 6324878 12 pageshttpsdoiorg10115520196324878

number of extreme precipitation days the amount of ex-treme precipitation and its percentage in the total amount ofMeiyu rainfall all showed a significant upward trend Wanget al [14] indicated that the probability of extreme pre-cipitation from May to September increases in 1961ndash2011 inmost areas of the Yangtze-Huai regions Rong et al [15]demonstrated that extreme precipitation is stronger in thesemihumid region of the HRB while it is less extreme in thehumid zone

Frequency analysis reveals the statistical characteristicsof hydrological stochastic series and gives quantitativeprediction of possible long-term changes of hydrologicalseries in the sense of probability (eg return levels andreturn periods) +erefore frequency analysis can providesome important information about the risk of flood disastersin a region However in hydrological frequency analysis ifthe hydrological series are nonstationary (eg with signif-icant trends or abrupt changes) the nonstationarity will havean important impact on the estimated return levels of ex-treme precipitation and runoff further influencing the as-sessment of flood risk [16 17] Under the background ofglobal change the stationarity of hydrological series has beendestroyed to a certain degree [18] and we will face morenonstationarity in the future How the nonstationarity inextreme precipitation and runoff series exert impacts on theestimation of return levels +is problem still needs to beinvestigated

In the context of the interdecadal changes of the EastAsian summer monsoon there have been many north-southmigrations in the rain belt in eastern China since the 1950s[19] For example in the 1980s the rain belt was mainlylocated in the middle and lower reaches of the Yangtze Riverand it has also experienced an interdecadal northward shiftafter the 1990s showing an increase in summer precipitationin the HRB [20] In this context it is very important toreexamine the characteristics of extreme precipitation in theHRB using updated precipitation data Floods and droughtsare more closely related to extreme precipitation over longerperiods of time (eg multiday extreme precipitation) whichhas not been studied sufficiently so far In this paper thetemporal and spatial changes and frequency analysis of 5-day and 10-day extreme precipitations in the HRB are ex-plored by trend analysis ARCGIS spatial analysis EOFanalysis and hydrological frequency analysis

2 Data and Methods

21 Data +e data in this paper are derived from theprecipitation observation data set provided by the NationalMeteorological Information Center +e National Meteo-rological Information Center has carried out strict qualitycontrol of the data such as correcting some suspiciousfalseobservations and correcting some data heterogeneity thatmay be caused by site migration and observation equipmentupgrades +e data selected in this paper are the dailyprecipitation data of 47 sites in the HRB from 1960 to 2014+e European Commission-funded project Statistical andRegional Dynamical Downscaling of Extremes for EuropeanRegions (STARDEX) was carried out to enhance the

methods for downscaling extreme temperature and pre-cipitation fromGCM [21 22] Floods and droughts are moreclosely related to extreme precipitation over longer periodsof time +erefore the annual maximum 5-day and 10-dayprecipitations namely the indices of PX5D and PX10D areselected in this study +ey are calculated by using thesoftware of STARDEX +e location of the selected stationsis shown in Figure 1

22 Methods

221 Empirical Orthogonal Analysis +e empirical or-thogonal function (EOF) method has obvious advantages inextracting information of spatiotemporal changes of phys-ical fields and has been an important method for meteo-rologists to analyze data [23 24]+e EOF analysis method isas follows all the data are arranged into an X matrix of nrows and m columns where n is the number of meteoro-logical stations and m is the observations of meteorologicalelements (each row represents the m observations of ameteorological station) +e EOF method is to decomposethe spatiotemporal data X into time vectors Z and spacevectors V

X VZ (1)

where V is a matrix of n row and n column each columnrepresents a typical space field Z is an n-row andm-columnmatrix and each row represents the time-weighted co-efficients relevant to corresponding column in V +e spacevectors are orthogonal to each other

VT

middot V I (2)

where I is a unit matrix +e time vectors are alsoorthogonal

Z middot ZT

Λ (3)

where Λ is a diagonal matrix V and Z can be derived fromthe matrix X after defining the matrix A as follows

A X middot XT (4)

+e data will be processed as departure from the meanvalues of each station +erefore A is a covariance matrixand also a real matrix According to the factorization the-orem of real symmetric matrices we have

A VΛVT (5)

where V is the eigenvector of A and also the space vector ofX andΛ is the diagonal matrix ofAwhose principal diagonalis the eigenvalue of A and the rest are all 0 +en the timevectors Z can be obtained as

Z VTX (6)

222 Hydrological Frequency Analysis In the hydrologicalfrequency analysis the linear moment method (L-moment)proposed by Hosking [25 26] is adopted in the parameter

2 Advances in Meteorology

estimation and the unbiased sample moment estimationmethod is used in the calculation

Let X1n leX2n le middot middot middot leXnn be the order statistics of arandom sample of size n which is drawn from the distri-bution of X Define the rth L-moment of variable X to be

λr equiv rminus1

1113944

rminus1

k0(minus1)

krminus 1

k1113888 1113889EXrminuskr r 1 2 (7)

where EXrminuskr is the (rminus k) th order statistics from a samplesize of r

+e Weibull formula is used as the plotting positionformula which is given as [27]

Pm m

n + 1 (8)

where n is the sample size and m is the rank of the ob-servations in the ascending order

Candidate fitting distribution types include generalizedextreme value distribution (GEV) Pearson type III distri-bution (P-III) three-parameter lognormal distribution(LN3) generalized logistic distribution (GLOG) distribu-tion value I-type distribution (EV1) generalized Paretodistribution (GPAR) Weibull distribution (Weibul) normaldistribution (N) exponential distribution (EXP) logisticdistribution (LOG) two-parameter lognormal distribution(LN2) gamma distribution (GAM) logarithmic Pearsontype III distribution (LP-III) four-parameterWakeby (wk4)and five-parameter Wakeby (wk5) +e chi-square andKolmogorovndashSmirnov tests were used for the selection of

distributions which are described in detail by Rao andHamed [28]

3 Results and Analysis

31 Temporal Changes of PX5D and PX10D Figure 2 ex-hibits the MannndashKendall trend of the PX5D and PX10Dseries during 1960ndash2014 in the HRB For PX5D more thanhalf of the stations have positive trend but only one station(Xihua Station) is significant at the 005 level and the other(Wuhe Station) is significant at the 01 significance level ForPX10D the results are quite similar most of the stationshave positive trends with two stations (Xihua Station andWuhe Station) significant at the 01 level No stations withnegative trend are significant at 005 or 01 level for the twoindices +erefore in general the PX5D and PX10D indiceshave a weak upward trend in the HRB

+e average PX5D and PX10D in the basin were quitesimilar to each other from 1960 to 2014 (Figure 3) And thecorrelation coefficient between them is up to 095 +e linearfitting curves of them have an upward trend according to theregression coefficients but not significant +e MannndashKendall trend test also showed that there were no significanttrends for them However linear fitting curves of the twoindices after the year 2003 indicated that the downwardtrend of them during this period both reached a significancelevel of 005 +erefore the significant decline of extremeprecipitation in the twenty-first century leads to the overallincrease of the two series to be nonsignificant

Combined with the flood and drought damaged areas inthe HRB from 1960 to 2000 it can be found that there is

Xihua

Shangqiu

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

LixinTaihe

Xinyi

Yinan

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

GaoyouFuyang

Bozhou

Bengbu

Xinghua

Baoying

Huaibei

Shenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

FengxianDengfeng

Shouxian

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou

114degE 116degE 118degE 120degE112degE

36degN

34degN

32degN

30degN0 100 20050

Kilometers

N

Basin boundary

Important stationsStations

Figure 1 Distribution of meteorological stations in the Huai River basin (HRB) Xihua station and Shangqiu station are labled asldquoimportant stationsrdquo which will be used in the frequency analysis in Section 33

Advances in Meteorology 3

good correspondence between the peaksvalleys of PX5DPX10D and the floodsdroughts in the basin (Figure 4) +epeaks and valleys of PX5D and PX10D are consistent withthose of the flood areas in Figures 4(a) and 4(c) while thepeaks and valleys of PX5D and PX10D are in reverse to those

of the drought areas in Figures 4(b) and 4(d)+e correlationcoefficients between PX5D (PX10D) and flood and droughtdamaged areas are 050 (059) and minus038 (minus047) during theperiod from 1960 to 2000 respectively all of which reaching005 significant level PX5D and PX10D were mainly at their

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou

112degE 114degE 116degE 118degE 120degE

30degN

32degN

34degN

36degN

0 100 20050Kilometers

Positive 005 levelPositive 01 levelNegative nonsignificant

Positive nonsignificant

N

(a)

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


30degN

32degN

34degN

36degN




N

(b)

Figure 2 Spatial pattern of the MannndashKendall trend of extreme precipitation indices of (a) PX5D and (b) PX10D in the HRB


peaks in the years when floods occurred (eg 1991 2003and 2007) and they were mainly at their valleys during theyears with severe droughts (eg 1992 1994 1999 and 2001)+erefore the changes of extreme precipitation are one of themajor factors influencing floods and droughts in the HRB

32 Spatiotemporal Patterns of PX5D and PX10D Empiricalorthogonal function (EOF) analysis is used to decomposethe extreme precipitation into spatial and associated tem-poral patterns +e North test proposed by North et al [29]

is used to test whether the modes of EOF analysis are sig-nificant and the results indicate that the first two modes ofPX5D passed the North test EOF results of PX5D are shownin Figure 5 and the results of PX10D (not shown) are verysimilar to PX5D Figures 5(a) and 5(b) show the leadingmodes of PX5D in the HRB EOF1 accounts for 2093 ofthe total variance From 5(a) it is found that the first modehas a loading which is almost entirely positive in the basinand only some individual sites in the northwest are negativereflecting the spatial consistency of PX5D Besides largevalues are concentrated over the mid- and southern parts of

PX5Dy = 00873x ndash 20252

P = 0711

PX10Dy = 00618x + 69651

P = 0851

PX5D-1y = ndash92137x + 18712

P = 0019

PX10D-1y = ndash6308x + 12833

P = 00150

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010Pr

ecip

itatio

n (m

m)

Year

PX5DPX10D

PX5D-1PX10D-1

Figure 3 Changes of extreme precipitation indices of PX5D and PX10D in the HRB Dashed lines indicate the linear trend of PX5D andPX10D and black and green lines indicate the trends of PX5D and PX10D during the period between 2003 and 2014

ndash2

ndash1

0

1

2

3

4

5

1960 1965 1970 1975 1980 1985 1990 1995 2000

PX5DFlood damaged areas

(a)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000

PX5DDrought damaged areas

(b)

ndash2

ndash1

0

1

2

3

4

5


1960 1965 1970 1975 1980 1985 1990 1995 2000

(c)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(d)

Figure 4 Relation between PX5DPX10D and flooddrought damaged areas in the HRB (standardized series) (a c) flood damaged areas(b d) drought damaged areas


the basin +e first principal component (PC1) has a weakupward trend but not significant (Figure 5(b)) and this issimilar to the trend of PX5D in Section 31 Combined withEOF1 1990 2001 and 2010 are the three years with thehighest PX5D and 1976 1978 and 1993 are the years withthe lowest PX5D

Figures 5(c) and 5(d) are the EOF2 and the secondprincipal component (PC2) of the PX5D +e percentage ofvariance explained by EOF2 is 1121 It can be found from5(c) that there is a significant spatial difference between thenorth and the south +e northern part of the basin isdominated by negative values and the southern part ispositive reflecting a meridional dipole pattern of PX5D+edownward trend of PC2 in Figure 5(d) is very weak

Semenov and Bengtsson [30] pointed out that globalwarming will accelerate the water cycle change the in-tensity and spatiotemporal distributions of precipitationand increase the occurrence possibility of regional extremeprecipitation events PC1 of PX5D shows an upward trendand this also confirms the viewpoint Zhu et al [31]demonstrated that the second eigenvector of precipitationin the Meiyu period exhibits the meridional dipole patternin the region which is consistent with the pattern of EOF2in this paper +eir study also shows that the strength of

East Asian subtropical summer monsoon the position ofwest Pacific subtropical high and the 200 hPa South Asianhigh are the main factors affecting the meridional dipolepattern And the previous winter ENSO also has a certainimpact on this kind of precipitation distribution Furtherresearch still needs to be done to explore the mechanismsof the spatiotemporal patterns of extreme precipitation inthe HRB

33 Frequency Analysis of PX5D and PX10D In this sectionthe effects of nonstationarity on the results of hydrologicalfrequency analysis are explored utilizing the typical series ofPX5D and PX10D with and without significant trends andabrupt changes in the HRB

+e MannndashKendall trend and abrupt change test areused to detect the stationarity of the PX5D and PX10D seriesin the HRB and most of the series do not have significanttrend or abrupt changes +e trend test results of PX5D andPX10D are shown in Figure 3 +e results of the abruptchange test of PX5D in Xihua Station and PX10D inShangqiu are shown in Figure 6 As can be seen PX5D wasmainly in the negative phase from 1960s to 1990s and was inthe positive phase after that reaching the 005 significance

021

015


31degN

32degN

33degN

34degN

35degN

36degN PX5D024

021

018

015

012

009

006

003

0

ndash003

009

015

009

021

015

003

(a)

y = 00172x ndash 34224P = 0534

ndash6

ndash4

ndash2

0

2

4

6

8

1960 1970 1980 1990 2000 2010

(b)

PX5D

ndash012


31degN

32degN

33degN

34degN

ndash004

ndash012

004012

02

35degN

012

02

36degN 03202802402

0040160120080040ndash004ndash008ndash012ndash016ndash02

(c)

y = ndash00003x + 05814P = 0988

ndash8

ndash6

ndash4

ndash2

0

2

4

6

1960 1970 1980 1990 2000 2010

(d)

Figure 5 EOF results of PX5D in HRB (a) EOF1 (b) PC1 (c) EOF2 (d) PC2


1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010minus5minus4minus3minus2minus1

012345

Year

Stat

istic

s

UF statisticsUB statistics005 significance level

(a)



012345

Year

Stat

istic

s

(b)

Figure 6 e MannndashKendall abrupt change test of PX5D in Xihua (a) and PX10D in Shangqiu (b)

y = 10077x + 11337P = 0049

0

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010

(a)

y = 051x + 12036P = 0299

0

50

100

150

200

250

300

350

400

450

1960 1970 1980 1990 2000 2010

(b)

Figure 7 Time series of PX5D in Xihua (a) and PX10D in Shangqiu (b)e red dashed line shows the means of the subseries and the blackdashed lines show the linear trends

0

100

200

300

400

500

600

Qua

ntile

Distribution comparison

P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

Figure 8 Continued


level in the middle and late 2000s Judging from the UF andUB curves the abrupt change occurred around 1991 +e Zstatistics of theMannndashKendall trend test for the total series is217 thus the upward trend reached the 005 significancelevel In a word there is a significant trend and abruptchange in PX5D of Xihua Station indicating that the seriesare nonstationary In Figure 6(b) there is no abrupt changein PX10D of Shangqiu Station and the Z statistics is only054 thus the PX10D series can be considered as stationary

+e original time series of PX5D in Xihua and PX10D inShangqiu are shown in Figure 7 As can be seen the linearupward trend of PX5D in Xihua is significant at the 005level and themean values of the two subtime series (between

1960 and 1991 and between 1992 and 2014) are quite dif-ferent +e average PX5D was low before 1991 but it washigher after 1991 And the PX10D series in Shangqiu arequite stationary with no significant trend (but the year 2012has an outlier)

+e frequency analysis of PX5D in Xihua Station from1960 to 2014 was first carried out According to thegoodness of fit test GEV distribution was selected as thebest distribution (at the 5 significance level based on bothchi-square test and KolmogorovndashSmirnov test) Figure 8(a)gives the fitting result of GEV and compares the fittingcurve of GEV with those of P-III and GPAR It can be seenthat the fitting of GEV distribution is better than the other

0

100

200

300

400

500

600

Qua

ntile

1960minus2014 GEV distribution

P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

TheoretRegional Observed

95 CI

(b)

Figure 8 (a) Comparison of fitting curves of GEV P-III and GPAR (b) Observed and estimated PX5D in Xihua station based on regionaland at-site methods

Series 1Series 2Series 3

0

100

200

300

400

500

600

700

PX5D

(mm

)

10 1001 1000Return periods (yr)

Figure 9 Frequency curves of the whole series (series 1) before (series 2) and after (series 3) the abrupt changes series of PX5D in Xihuastation


two distributions +e observed and estimated PX5D basedon regional and at-site methods is shown in Figure 8(b)+e red line is the regional frequency curve (all the 47stations in the study area are used as one homogeneousregion) +e blue line is the theoretical frequency curvewhich is based on the at-site method And the red line andthe blue line are very close to each other It shows that allthe observed points are close to the GEV curve and theestimated values at the 5ndash20-year return periods areunderestimated but all the observed points are within the95 confidence interval of the theoretical frequency curve+e GEV curve in Figure 8(b) is approximately a linewhich is quite different from the GEV curve in Figure 8(a)and this is because the spacing of horizontal ordinate haschanged in Figure 8(a)

PX5D series of Xihua Station was further divided intotwo subseries series 2 (1961ndash1991) is before the abruptchange and series 3 (1992ndash2014) is after the abrupt changeAnd the whole series (1960sim2014) is labeled as series 1 GEVis the best distribution for both series 1 and 2 Figure 9 is theestimated quantiles of PX5D for the three series at differentreturn periods As can be seen the frequency curve of series2 is significantly higher than that of series 1 and the curve ofseries 3 is located between the former two Meanwhile thedifferences between the estimation of series 1 and 2 in-creased with the increase of return periods +erefore theestimated quantiles (return levels) are small and unsafe whenthe analysis is based on series 1 or series 3 In addition theHurst index calculated by the RS method is 091 for series 3which indicates that PX5D of Xihua Station will increase in

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)

Figure 10 (a) Comparison of fitting curves of GEV P-III and GPAR (b) Observed and estimated PX10D in Shangqiu Station based onregional and at-site methods


the future and the extreme precipitation estimated by thewhole series is even more unsafe

+e goodness of fit test indicated that GEV distribu-tion is also the best distribution for PX10D in ShangqiuStation (at the 5 significance level for both chi-squareand KolmogorovndashSmirnov test) Figure 10(a) gives thefitting result of GEV and compares the fitting curve ofGEV with those of P-III and GPAR +e observed andestimated PX10D based on regional and at-site methods isshown in Figure 10(b) Besides both Figures 10(a) and10(b) have an outlier which occurred in 2012 But all theobserved points including the outlier are within the 95confidence interval of the theoretical frequency curve inFigure 10(b)

Figure 11 shows the estimated quantiles (return levels)of PX10D in Shangqiu Station at different return periods+e return level of 100-year-PX10D is about 3628 mmand that of 50-year-PX10D is about 3266 mm +emaximum PX10D in Shangqiu was 4122 mm in 2012which is up to a return period of nearly 300 years +eyear with the second biggest PX10D is 2004 (PX10D isonly 2885 mm) and the return period is less than30 years

In the calculation of the quantiles of extreme pre-cipitation if the whole series is directly utilized in the fre-quency analysis and the stationarity test of the data series isneglected two kinds of problems will occur on the onehand it does not meet the basic assumptions on data sta-tionarity in frequency analysis on the other hand the es-timated quantile is significantly different from those basedon the series before or after the abrupt change and this effectis more significant for longer return periods +e estimatedextreme precipitation calculated according to the series afterthe abrupt change has higher accuracy but this also in-creases the uncertainty of parameter estimation since thelength (sample size) is significantly smaller than the wholeseries In the future when extreme precipitation continuesto increase the estimation by the whole series will be evenmore unsafe Salas and Obeysekera [32] proposed a simpleand unified framework to estimate the return period and riskfor nonstationary hydrologic events Gilleland and Katz [33]

also proposed a GEV model based on nonstationary seriesrecently +e estimated extreme precipitation changes withtime and is known as ldquoeffective return levelrdquo However howto apply the newmethods to engineering hydrological designremains to be an open question [34 35] +erefore fre-quency analysis of the nonstationary series needs to bestudied in depth to obtain better estimation of extremeprecipitation

4 Conclusions

Floods and droughts are more closely related to extremeprecipitation over longer periods of time +e spatial andtemporal changes and frequency analysis of 5-day and10-day extreme precipitations in the Huai River basin(HRB) are investigated by means of correlation analysistrend and abrupt change analysis EOF analysis andhydrological frequency analysis based on the daily pre-cipitation data from 1960 to 2014

Generally more stations have positive trends for PX5Dand PX10D in the HRB PX5D and PX10D indices have aweak upward trend in the HRB and PC1 of the EOF analysisalso has a weak upward trend +e weak upward trend maymainly be due to the significant downward trend in the 21stcentury

+e changes of PX5D and PX10D are well consistentwith the flood and drought damaged areas in the basin +isindicates that the multiday (5-day and 10-day) extremeprecipitation is closely associated with flooddrought di-sasters and can be used to study the risk of flood and droughtin the future

When the hydrological time series satisfy the stationarityconsumption of frequency analysis the series can be usedwithout any changes Nonstationary hydrological series withsignificant trends or abrupt changes will have importantimpacts in frequency analysis For the stations with sig-nificant upward trend and abrupt changes if the whole seriesare used for frequency analysis the estimated quantiles willsignificantly be lower thus the risk of future flooding will beunderestimated and this effect will be more pronounced forlonger return periods

0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)

Figure 11 Frequency curve of PX10D in Shangqiu Station


Data Availability

+e data used to support the findings of this study areavailable from the corresponding author upon request

Conflicts of Interest

+e authors declare that they have no conflicts of interest

Acknowledgments

+is study was supported by the National Natural ScienceFoundation of China (41671022 and 41575094) Young Top-Notch Talent Support Program of National High-levelTalents Special Support Plan Strategic Consulting Projectsof Chinese Academy of Engineering (2016-ZD-08-05-02)and Meteorological Research Foundation of the Huai RiverBasin (HRM201701)

References

[1] L V Alexander X Zhang T C Peterson et al ldquoGlobalobserved changes in daily climate extremes of temperatureand precipitationrdquo Journal of Geophysical Research vol 111no 5 article D05109 2006

[2] Y Chen and P M Zhai ldquoPersistent extreme precipitationevents in China during 1951ndash2010rdquo Climate Research vol 57no 2 pp 143ndash155 2013

[3] C Onyutha ldquoGeospatial trends and decadal anomalies inextreme rainfall over Uganda East Africardquo Advances inMeteorology vol 2016 Article ID 6935912 15 pages 2016

[4] Z X Zhang Q Jin X Chen C-Y Xu and S S Jiang ldquoOn thelinkage between the extreme drought and pluvial patterns inChina and the large-scale atmospheric circulationrdquo Advancesin Meteorology vol 2016 Article ID 8010638 12 pages 2016

[5] C R Balling Jr M S K Kiany S S Roy and J KhoshhalldquoTrends in extreme precipitation indices in Iran 1951ndash2007rdquoAdvances in Meteorology vol 2016 Article ID 24568098 pages 2016

[6] P Zhai X Zhang H Wan and X Pan ldquoTrends in totalprecipitation and frequency of daily precipitation extremesover Chinardquo Journal of Climate vol 18 no 7 pp 1096ndash11082005

[7] YWang Q Zhang and V P Singh ldquoSpatiotemporal patternsof precipitation regimes in the Huai River basin China andpossible relations with ENSO eventsrdquo Natural Hazardsvol 82 no 3 pp 2167ndash2185 2016

[8] G Y An and Z C Hao ldquoVariation of precipitation andstreamflow in the upper and middle Huaihe river basinChina from 1959ndash2009rdquo Journal of Coastal Research vol 80pp 69ndash79 2017

[9] Z Pan X Ruan M Qian J Hua N Shan and J Xu ldquoSpatio-temporal variability of streamflow in the Huaihe River BasinChina climate variability or human activitiesrdquo HydrologyResearch vol 49 no 1 pp 177ndash193 2018

[10] Z W Ye and Z H Li ldquoSpatiotemporal variability and trendsof extreme precipitation in the Huaihe River Basin a climatictransitional zone in East Chinardquo Advances in Meteorologyvol 2017 Article ID 3197435 15 pages 2017

[11] J Xia D She Y Zhang and H Du ldquoSpatio-temporal trendand statistical distribution of extreme precipitation events inHuaihe River Basin during 1960-2009rdquo Journal of Geo-graphical Sciences vol 22 no 2 pp 195ndash208 2012

[12] Z G Lu X H Zhang J L Huo K Q Wang and X P Xieldquo+e evolution characteristics of the extreme precipitation inHuaihe river basin during 1960-2008rdquo Journal of the Mete-orological Sciences vol 31 no 1 pp 74ndash80 2011 in Chinese

[13] W Yang and Z Cheng ldquoVariation characteristics of extremeprecipitation during Meiyu flood period over the Yangtze-Huaihe basin in recent 53 yearsrdquo Meteorological Monthlyvol 41 no 9 pp 1126ndash1133 2015 in Chinese

[14] J Wang J H Yu and J Q He ldquoStudy on characteristics andchange trend of extreme rainfall in the Jianghuai regionrdquoClimatic and Environmental Research vol 20 no 1 pp 80ndash88 2015 in Chinese

[15] Y S Rong W Wang P Wang and L Y Bai ldquoAnalysis ofcharacteristics of extreme rainfall and estimate of rainfallduring return periods in Huaihe Basinrdquo Journal of HohaiUniversity (Natural Sciences) vol 40 no 1 pp 1ndash8 2012 inChinese

[16] G Villarini J A Smith F Serinaldi J Bales P D Bates andW F Krajewski ldquoFlood frequency analysis for nonstationaryannual peak records in an urban drainage basinrdquo Advances inWater Resources vol 32 no 8 pp 1255ndash1266 2009

[17] A T Silva M M Portela and M Naghettini ldquoNon-stationarities in the occurrence rates of flood events in Por-tuguese watershedsrdquo Hydrology and Earth System Sciencesvol 16 no 1 pp 241ndash254 2012

[18] P C D Milly J Betancourt M Falkenmark et al ldquoStatio-narity is dead whither water managementrdquo Science vol 319no 5863 pp 573-574 2008

[19] X Y Lv X Z Zhang and J N Chen ldquo+e interdecadalvariation of advance and retreat of east Asian summermonsoon and their effect on the regional rainfall over ChinardquoJournal of Tropical Meteorology vol 19 no 4 pp 340ndash3482013

[20] H Liu T Zhou Y Zhu and Y Lin ldquo+e strengthening EastAsia summer monsoon since the early 1990srdquo Chinese ScienceBulletin vol 57 no 13 pp 1553ndash1558 2012

[21] M R Haylock and C M Goodess ldquoInterannual variability ofEuropean extreme winter rainfall and links with mean large-scale circulationrdquo International Journal of Climatologyvol 24 no 6 pp 759ndash776 2004

[22] A Moberg and P D Jones ldquoTrends in indices for extremes indaily temperature and precipitation in central and westernEurope 1901-99rdquo International Journal of Climatologyvol 25 no 9 pp 1149ndash1171 2005

[23] C Svensson ldquoEmpirical orthogonal function analysis of dailyrainfall in the upper reaches of the Huai River basin Chinardquogteoretical and Applied Climatology vol 62 no 3-4pp 147ndash161 1999

[24] A K Smith and H M F Semazzi ldquo+e role of the dominantmodes of precipitation variability over Eastern Africa inmodulating the hydrology of lake victoriardquo Advances inMeteorology vol 2014 Article ID 516762 11 pages 2014

[25] J R M Hosking ldquoL-moments analysis and estimation ofdistributions using linear combinations of order statisticsrdquoJournal of the Royal Statistical Society Series B (Methodo-logical) vol 52 no 1 pp 105ndash124 1990

[26] J R M Hosking and J R Wallis Regional Frequency AnalysisAn Approach Based on L-Moments Cambridge UniversityPress Cambridge UK 1997

[27] C Cunnane ldquoStatistical distributions for flood frequencyanalysisrdquo World Meteorological Orgnization OperationalHydrology Report No 33 WMO-No 718 WMO GenevaSwitzerland 1989


[28] A R Rao and K H Hamed Flood Frequency Analysis CRCPress LLC Boca Raton FL USA 2000

[29] G R North T L Bell R F Cahalan and F J MoengldquoSampling errors in the estimation of empirical orthogonalfunctionsrdquo Monthly Weather Review vol 110 no 7pp 699ndash706 1982

[30] V Semenov and L Bengtsson ldquoSecular trends in daily pre-cipitation characteristics greenhouse gas simulation with acoupled AOGCMrdquo Climate Dynamics vol 19 no 2pp 123ndash140 2002

[31] X Zhu J He and ZWu ldquoMeridional seesaw-like distributionof theMeiyu rainfall over the Changjiang-Huaihe River Valleyand characteristics in the anomalous climate yearsrdquo ChineseScience Bulletin vol 52 no 17 pp 2420ndash2428 2007

[32] J D Salas and J Obeysekera ldquoRevisiting the concepts ofreturn period and risk for nonstationary hydrologic extremeeventsrdquo Journal of Hydrologic Engineering vol 19 no 3pp 554ndash568 2014

[33] E Gilleland and R W Katz ldquoextRemes 20 an extreme valueanalysis package in Rrdquo Journal of Statistical Software vol 72no 8 pp 1ndash39 2016

[34] L H Xiong C Jiang T Du S L Guo and C-Y Xu ldquoReviewon nonstationary hydrological frequency analysis underchanging environmentsrdquo Journal of Water Resources Re-search vol 4 no 4 pp 310ndash319 2015 in Chinese

[35] A Cancelliere ldquoNon stationary analysis of extreme eventsrdquoWater Resources Management vol 31 no 10 pp 3097ndash31102017


Hindawiwwwhindawicom Volume 2018

Journal of

ChemistryArchaeaHindawiwwwhindawicom Volume 2018

Marine BiologyJournal of


BiodiversityInternational Journal of


EcologyInternational Journal of


Hindawiwwwhindawicom

Applied ampEnvironmentalSoil Science

Volume 2018

Forestry ResearchInternational Journal of



International Journal of

Geophysics

Environmental and Public Health

Journal of




Microbiology


Public Health Advances in

AgricultureAdvances in


Agronomy




MeteorologyAdvances in

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018Hindawiwwwhindawicom Volume 2018

ChemistryAdvances in

ScienticaHindawiwwwhindawicom Volume 2018


Geological ResearchJournal of

Analytical ChemistryInternational Journal of


Submit your manuscripts atwwwhindawicom

number of extreme precipitation days the amount of ex-treme precipitation and its percentage in the total amount ofMeiyu rainfall all showed a significant upward trend Wanget al [14] indicated that the probability of extreme pre-cipitation from May to September increases in 1961ndash2011 inmost areas of the Yangtze-Huai regions Rong et al [15]demonstrated that extreme precipitation is stronger in thesemihumid region of the HRB while it is less extreme in thehumid zone

Frequency analysis reveals the statistical characteristicsof hydrological stochastic series and gives quantitativeprediction of possible long-term changes of hydrologicalseries in the sense of probability (eg return levels andreturn periods) +erefore frequency analysis can providesome important information about the risk of flood disastersin a region However in hydrological frequency analysis ifthe hydrological series are nonstationary (eg with signif-icant trends or abrupt changes) the nonstationarity will havean important impact on the estimated return levels of ex-treme precipitation and runoff further influencing the as-sessment of flood risk [16 17] Under the background ofglobal change the stationarity of hydrological series has beendestroyed to a certain degree [18] and we will face morenonstationarity in the future How the nonstationarity inextreme precipitation and runoff series exert impacts on theestimation of return levels +is problem still needs to beinvestigated

In the context of the interdecadal changes of the EastAsian summer monsoon there have been many north-southmigrations in the rain belt in eastern China since the 1950s[19] For example in the 1980s the rain belt was mainlylocated in the middle and lower reaches of the Yangtze Riverand it has also experienced an interdecadal northward shiftafter the 1990s showing an increase in summer precipitationin the HRB [20] In this context it is very important toreexamine the characteristics of extreme precipitation in theHRB using updated precipitation data Floods and droughtsare more closely related to extreme precipitation over longerperiods of time (eg multiday extreme precipitation) whichhas not been studied sufficiently so far In this paper thetemporal and spatial changes and frequency analysis of 5-day and 10-day extreme precipitations in the HRB are ex-plored by trend analysis ARCGIS spatial analysis EOFanalysis and hydrological frequency analysis

2 Data and Methods

21 Data +e data in this paper are derived from theprecipitation observation data set provided by the NationalMeteorological Information Center +e National Meteo-rological Information Center has carried out strict qualitycontrol of the data such as correcting some suspiciousfalseobservations and correcting some data heterogeneity thatmay be caused by site migration and observation equipmentupgrades +e data selected in this paper are the dailyprecipitation data of 47 sites in the HRB from 1960 to 2014+e European Commission-funded project Statistical andRegional Dynamical Downscaling of Extremes for EuropeanRegions (STARDEX) was carried out to enhance the

methods for downscaling extreme temperature and pre-cipitation fromGCM [21 22] Floods and droughts are moreclosely related to extreme precipitation over longer periodsof time +erefore the annual maximum 5-day and 10-dayprecipitations namely the indices of PX5D and PX10D areselected in this study +ey are calculated by using thesoftware of STARDEX +e location of the selected stationsis shown in Figure 1

22 Methods

221 Empirical Orthogonal Analysis +e empirical or-thogonal function (EOF) method has obvious advantages inextracting information of spatiotemporal changes of phys-ical fields and has been an important method for meteo-rologists to analyze data [23 24]+e EOF analysis method isas follows all the data are arranged into an X matrix of nrows and m columns where n is the number of meteoro-logical stations and m is the observations of meteorologicalelements (each row represents the m observations of ameteorological station) +e EOF method is to decomposethe spatiotemporal data X into time vectors Z and spacevectors V

X VZ (1)

where V is a matrix of n row and n column each columnrepresents a typical space field Z is an n-row andm-columnmatrix and each row represents the time-weighted co-efficients relevant to corresponding column in V +e spacevectors are orthogonal to each other

VT

middot V I (2)

where I is a unit matrix +e time vectors are alsoorthogonal

Z middot ZT

Λ (3)

where Λ is a diagonal matrix V and Z can be derived fromthe matrix X after defining the matrix A as follows

A X middot XT (4)

+e data will be processed as departure from the meanvalues of each station +erefore A is a covariance matrixand also a real matrix According to the factorization the-orem of real symmetric matrices we have

A VΛVT (5)

where V is the eigenvector of A and also the space vector ofX andΛ is the diagonal matrix ofAwhose principal diagonalis the eigenvalue of A and the rest are all 0 +en the timevectors Z can be obtained as

Z VTX (6)

222 Hydrological Frequency Analysis In the hydrologicalfrequency analysis the linear moment method (L-moment)proposed by Hosking [25 26] is adopted in the parameter




λr equiv rminus1

1113944

rminus1

k0(minus1)

krminus 1

k1113888 1113889EXrminuskr r 1 2 (7)



Pm m

n + 1 (8)








Xihua

Shangqiu

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

LixinTaihe

Xinyi

Yinan

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

GaoyouFuyang

Bozhou

Bengbu

Xinghua

Baoying

Huaibei

Shenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

FengxianDengfeng

Shouxian

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


36degN

34degN

32degN

30degN0 100 20050

Kilometers

N

Basin boundary






Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


30degN

32degN

34degN

36degN




N

(a)

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


30degN

32degN

34degN

36degN




N

(b)






PX5Dy = 00873x ndash 20252

P = 0711

PX10Dy = 00618x + 69651

P = 0851

PX5D-1y = ndash92137x + 18712

P = 0019

PX10D-1y = ndash6308x + 12833

P = 00150

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010Pr

ecip

itatio

n (m

m)

Year

PX5DPX10D

PX5D-1PX10D-1


ndash2

ndash1

0

1

2

3

4

5

1960 1965 1970 1975 1980 1985 1990 1995 2000


(a)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(b)

ndash2

ndash1

0

1

2

3

4

5


1960 1965 1970 1975 1980 1985 1990 1995 2000

(c)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(d)









021

015


31degN

32degN

33degN

34degN

35degN

36degN PX5D024

021

018

015

012

009

006

003

0

ndash003

009

015

009

021

015

003

(a)

y = 00172x ndash 34224P = 0534

ndash6

ndash4

ndash2

0

2

4

6

8

1960 1970 1980 1990 2000 2010

(b)

PX5D

ndash012


31degN

32degN

33degN

34degN

ndash004

ndash012

004012

02

35degN

012

02

36degN 03202802402


(c)

y = ndash00003x + 05814P = 0988

ndash8

ndash6

ndash4

ndash2

0

2

4

6

1960 1970 1980 1990 2000 2010

(d)




012345

Year

Stat

istic

s


(a)



012345

Year

Stat

istic

s

(b)


y = 10077x + 11337P = 0049

0

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010

(a)

y = 051x + 12036P = 0299

0

50

100

150

200

250

300

350

400

450

1960 1970 1980 1990 2000 2010

(b)


0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

Figure 8 Continued






0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)



0

100

200

300

400

500

600

700

PX5D

(mm

)






0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy

















λr equiv rminus1

1113944

rminus1

k0(minus1)

krminus 1

k1113888 1113889EXrminuskr r 1 2 (7)



Pm m

n + 1 (8)








Xihua

Shangqiu

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

LixinTaihe

Xinyi

Yinan

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

GaoyouFuyang

Bozhou

Bengbu

Xinghua

Baoying

Huaibei

Shenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

FengxianDengfeng

Shouxian

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


36degN

34degN

32degN

30degN0 100 20050

Kilometers

N

Basin boundary






Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


30degN

32degN

34degN

36degN




N

(a)

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


30degN

32degN

34degN

36degN




N

(b)






PX5Dy = 00873x ndash 20252

P = 0711

PX10Dy = 00618x + 69651

P = 0851

PX5D-1y = ndash92137x + 18712

P = 0019

PX10D-1y = ndash6308x + 12833

P = 00150

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010Pr

ecip

itatio

n (m

m)

Year

PX5DPX10D

PX5D-1PX10D-1


ndash2

ndash1

0

1

2

3

4

5

1960 1965 1970 1975 1980 1985 1990 1995 2000


(a)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(b)

ndash2

ndash1

0

1

2

3

4

5


1960 1965 1970 1975 1980 1985 1990 1995 2000

(c)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(d)









021

015


31degN

32degN

33degN

34degN

35degN

36degN PX5D024

021

018

015

012

009

006

003

0

ndash003

009

015

009

021

015

003

(a)

y = 00172x ndash 34224P = 0534

ndash6

ndash4

ndash2

0

2

4

6

8

1960 1970 1980 1990 2000 2010

(b)

PX5D

ndash012


31degN

32degN

33degN

34degN

ndash004

ndash012

004012

02

35degN

012

02

36degN 03202802402


(c)

y = ndash00003x + 05814P = 0988

ndash8

ndash6

ndash4

ndash2

0

2

4

6

1960 1970 1980 1990 2000 2010

(d)




012345

Year

Stat

istic

s


(a)



012345

Year

Stat

istic

s

(b)


y = 10077x + 11337P = 0049

0

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010

(a)

y = 051x + 12036P = 0299

0

50

100

150

200

250

300

350

400

450

1960 1970 1980 1990 2000 2010

(b)


0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

Figure 8 Continued






0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)



0

100

200

300

400

500

600

700

PX5D

(mm

)






0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy

















Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


30degN

32degN

34degN

36degN




N

(a)

Suyu

Wuhe

Juye

Xuyi

Luan

Funan

Taihe

Xinyi

Yinan

Xihua

Ganyu

JianhuSixian

LankaoJining

Xuzhou

Suzhou

RizhaoJuxian

Gaoyou

Fuyang

Bozhou

BengbuXinghua

BaoyingShenqiu

GuanyunSuixian

Xinxian

Yanzhou

Xuchang

Xinyang

Sheyang

Kaifeng

Huoshan

Dongtai

Baofeng

Huaiyuan

Fengxian

Shouxian

Shangqiu

Huaiying

Dangshan

Yongcheng

Zaozhuang

Zhumadian

Zhengzhou


30degN

32degN

34degN

36degN




N

(b)






PX5Dy = 00873x ndash 20252

P = 0711

PX10Dy = 00618x + 69651

P = 0851

PX5D-1y = ndash92137x + 18712

P = 0019

PX10D-1y = ndash6308x + 12833

P = 00150

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010Pr

ecip

itatio

n (m

m)

Year

PX5DPX10D

PX5D-1PX10D-1


ndash2

ndash1

0

1

2

3

4

5

1960 1965 1970 1975 1980 1985 1990 1995 2000


(a)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(b)

ndash2

ndash1

0

1

2

3

4

5


1960 1965 1970 1975 1980 1985 1990 1995 2000

(c)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(d)









021

015


31degN

32degN

33degN

34degN

35degN

36degN PX5D024

021

018

015

012

009

006

003

0

ndash003

009

015

009

021

015

003

(a)

y = 00172x ndash 34224P = 0534

ndash6

ndash4

ndash2

0

2

4

6

8

1960 1970 1980 1990 2000 2010

(b)

PX5D

ndash012


31degN

32degN

33degN

34degN

ndash004

ndash012

004012

02

35degN

012

02

36degN 03202802402


(c)

y = ndash00003x + 05814P = 0988

ndash8

ndash6

ndash4

ndash2

0

2

4

6

1960 1970 1980 1990 2000 2010

(d)




012345

Year

Stat

istic

s


(a)



012345

Year

Stat

istic

s

(b)


y = 10077x + 11337P = 0049

0

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010

(a)

y = 051x + 12036P = 0299

0

50

100

150

200

250

300

350

400

450

1960 1970 1980 1990 2000 2010

(b)


0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

Figure 8 Continued






0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)



0

100

200

300

400

500

600

700

PX5D

(mm

)






0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy


















PX5Dy = 00873x ndash 20252

P = 0711

PX10Dy = 00618x + 69651

P = 0851

PX5D-1y = ndash92137x + 18712

P = 0019

PX10D-1y = ndash6308x + 12833

P = 00150

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010Pr

ecip

itatio

n (m

m)

Year

PX5DPX10D

PX5D-1PX10D-1


ndash2

ndash1

0

1

2

3

4

5

1960 1965 1970 1975 1980 1985 1990 1995 2000


(a)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(b)

ndash2

ndash1

0

1

2

3

4

5


1960 1965 1970 1975 1980 1985 1990 1995 2000

(c)

ndash2

ndash1

0

1

2

3

1960 1965 1970 1975 1980 1985 1990 1995 2000


(d)









021

015


31degN

32degN

33degN

34degN

35degN

36degN PX5D024

021

018

015

012

009

006

003

0

ndash003

009

015

009

021

015

003

(a)

y = 00172x ndash 34224P = 0534

ndash6

ndash4

ndash2

0

2

4

6

8

1960 1970 1980 1990 2000 2010

(b)

PX5D

ndash012


31degN

32degN

33degN

34degN

ndash004

ndash012

004012

02

35degN

012

02

36degN 03202802402


(c)

y = ndash00003x + 05814P = 0988

ndash8

ndash6

ndash4

ndash2

0

2

4

6

1960 1970 1980 1990 2000 2010

(d)




012345

Year

Stat

istic

s


(a)



012345

Year

Stat

istic

s

(b)


y = 10077x + 11337P = 0049

0

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010

(a)

y = 051x + 12036P = 0299

0

50

100

150

200

250

300

350

400

450

1960 1970 1980 1990 2000 2010

(b)


0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

Figure 8 Continued






0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)



0

100

200

300

400

500

600

700

PX5D

(mm

)






0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy





















021

015


31degN

32degN

33degN

34degN

35degN

36degN PX5D024

021

018

015

012

009

006

003

0

ndash003

009

015

009

021

015

003

(a)

y = 00172x ndash 34224P = 0534

ndash6

ndash4

ndash2

0

2

4

6

8

1960 1970 1980 1990 2000 2010

(b)

PX5D

ndash012


31degN

32degN

33degN

34degN

ndash004

ndash012

004012

02

35degN

012

02

36degN 03202802402


(c)

y = ndash00003x + 05814P = 0988

ndash8

ndash6

ndash4

ndash2

0

2

4

6

1960 1970 1980 1990 2000 2010

(d)




012345

Year

Stat

istic

s


(a)



012345

Year

Stat

istic

s

(b)


y = 10077x + 11337P = 0049

0

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010

(a)

y = 051x + 12036P = 0299

0

50

100

150

200

250

300

350

400

450

1960 1970 1980 1990 2000 2010

(b)


0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

Figure 8 Continued






0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)



0

100

200

300

400

500

600

700

PX5D

(mm

)






0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy
















012345

Year

Stat

istic

s


(a)



012345

Year

Stat

istic

s

(b)


y = 10077x + 11337P = 0049

0

50

100

150

200

250

300

350

1960 1970 1980 1990 2000 2010

(a)

y = 051x + 12036P = 0299

0

50

100

150

200

250

300

350

400

450

1960 1970 1980 1990 2000 2010

(b)


0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

Figure 8 Continued






0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)



0

100

200

300

400

500

600

700

PX5D

(mm

)






0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy



















0

100

200

300

400

500

600

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)



0

100

200

300

400

500

600

700

PX5D

(mm

)






0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy

















0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500

GEVP-III

GPARObserved

(a)

0

50

100

150

200

250

300

350

400

450

500

Qua

ntile


P ()T (yr)

1101

10111

502

805

9010

9520

9850

99100

995200

998500


95 CI

(b)








4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy




















4 Conclusions





0

100

200

300

400

500

600

1 10 100 1000

PX10

D (m

m)

Return periods (yr)



Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy















Data Availability




Acknowledgments


References







































Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy

























Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy
















Journal of










Volume 2018





Geophysics


Journal of




Microbiology





Agronomy















Documents

SpatiotemporalChangesandFrequencyAnalysisof …downloads.hindawi.com/journals/amete/2019/6324878.pdf · 2019. 8. 13. · O5qD O5λD O5PD O55D XDDD XDOD (b) j(– Fwfq ff5] ff(] ffX]