Changes of atmospheric water vapor budget in the Pearl

Changes of atmospheric water vapor budget in the Pearl
Theor Appl Climatol (2010) 102:185–195
DOI 10.1007/s00704-010-0257-z
ORIGINAL PAPER
Changes of atmospheric water vapor budget in the Pearl
River basin and possible implications for hydrological cycle
Qiang Zhang & Chong-Yu Xu & Zengxin Zhang &
Yongqin David Chen
Received: 1 March 2009 / Accepted: 11 January 2010 / Published online: 6 February 2010
# Springer-Verlag 2010
Abstract In this study, we thoroughly analyzed abrupt
behaviors, trends, and periodicity properties of water vapor
flux and moisture budget entering and exiting the four
edges of the Pearl River basin based on the NCAR/NCEP
reanalysis dataset by using the continuous wavelet transform and the simple two-phase linear regression technique.
Possible implications for hydrological cycle and water
resource management of these changes are also discussed.
The results indicate that: (1) the water vapor propagating
through the four edges of the Pearl River basin is
decreasing, and it is particularly true for the changes of
the water vapor flux exiting from the north edge of the
study river basin. The transition point from increase to
Q. Zhang (*)
Department of Water Resources and Environment,
Sun Yat-sen University,
Guangzhou 510275, China
e-mail: [email protected]
C.-Y. Xu
Department of Geosciences and Hydrology, University of Oslo,
Blindern, PO Box 1047, Oslo 0316, Norway
Z. Zhang
Jiangsu Key Laboratory of Forestry Ecological Engineering,
Nanjing Forestry University,
Nanjing 210037, China
Y. D. Chen
Department of Geography and Resource Management,
The Chinese University of Hong Kong,
Shatin, N.T.,
Hong Kong, China
Y. D. Chen
Centre of Strategic Environmental Assessment for China,
The Chinese University of Hong Kong,
Shatin, N.T.,
Hong Kong, China
decrease occurs in the early 1960s; (2) The wavelet
transform spectra indicate that the monthly water vapor
flux through the north edge decreases and this decrease is
mainly reflected by intermittent distribution of the wavelet
power spectra after early 1980s. The periodicity properties
of the water vapor flux through the north edge imply that
the northward propagation of water vapor flux decreases
after the 1980s; (3) close relations between water vapor
flux, precipitation and streamflow implies that the altered
hydrological cycle in the Pearl River basin is mainly
manifested by seasonal shifts of water vapor flux after
early 1960s. One of the direct consequences of these
changes of water vapor flux is the seasonal transition of wet
and dry conditions across the Pearl River basin. Regional
responses of hydrological cycle to climate variation/change
could be different from one river basin to another.
Hydrological responses of the Pearl River basin to the
global warming are mainly demonstrated by seasonal shifts
of precipitation changes: winter comes to be wetter and
summer tends to be dryer. The finding of the seasonal
transition of precipitation in the Pearl River basin is of great
scientific and practical merits in basin scale water resource
management in the Pearl River basin under the changing
climate and global warming in particular.
1 Introduction
Rising atmospheric carbon dioxide (CO2) contributes to
global warming and therefore to variations in both precipitation and evapotranspiration (ET; Goyal 2004; Kruijt et al.
2008). As for the water balance at regional scales, this raises
the question about whether CO2 increase will lead to increased
runoff or to increased water shortages. This concerns the
influences of global warming on the hydrological cycle at
186
global and regional scales. Hydrologists and meteorologists
suggested that an increase in surface temperature leads to
higher evaporation rates and enables the atmosphere to
transport higher amounts of water vapor, which, in turn,
leads to accelerated hydrological cycle (e.g. Menzel and
Bürger 2002). The accelerated hydrologic cycle of the last
two decades is believed to be one of the consequences of
global warming, especially in some parts of the northern
hemisphere (Brutsaert and Parlange 1998; Karl et al. 1996).
The present global warming has led to changes of the global
hydrological cycle and to the amplitude of the increase of
global and continental runoff (Semenov and Bengtsson 2002;
Labat et al. 2004; Milly et al. 2005). Gao et al. (2007)
identified different changing properties of hydrological cycle,
indicating that the hydrological cycle was intensified in
western China, whereas it was weakened from the Yellow
River basin northward. Chaplot (2007) demonstrated that
changes in CO2 concentration and climate, particularly an
increase in precipitation, had significant effects on the soil
and fresh water resources. However, the magnitude of these
effects is different according to the watershed’s characteristics. The increasing temperature leads to the changes of
atmospheric water budget due to the high sensitivity of the
saturation vapor pressure in air to temperature, and perturbations in the global water cycle are expected to accompany the
climate warming (Allen and Ingram 2002).
Precipitation efficiency is the fraction of the average
horizontal water vapor flux over an area that falls as rain.
Summer rainfall, whether forced by synoptic-scale disturbances or by mesoscale mechanisms, is overwhelmingly
subject to moisture transportation and deep convection
(Heideman and Fritsch 1988). Therefore, a lot of studies
attempted to address impacts of climate changes on
hydrological processes and underlying global circulation
of atmospheric water vapor flux and moisture budget (e.g.,
Thodsen 2007; Zhang et al. 2008d). Zhang et al. (2008a)
investigated spatial and temporal patterns of trends of
precipitation maxima in the Yangtze River basin during
1960–2005 and explored association of changing patterns
of the precipitation maxima with large-scale circulation
using NCEP/NCAR reanalysis data. They indicated that
decreasing strength of East Asian summer monsoon during
1975–2005 as compared to that during 1961–1974 and
increasing geopotential height in the north China, South
China Sea and west Pacific regions combine to negatively
impact the northward propagation of the water vapor flux.
This circulation pattern will be beneficial for the longer stay
of the Meiyu front in the Yangtze River basin, leading to
more precipitation in the middle and lower Yangtze River
basin in summer months. Zhang et al. (2008d) also
illustrated the consecutively increasing summer precipitation in the lower Yangtze River basin as a result of the
consistently increasing moisture budget.
Q. Zhang et al.
Similar results were obtained in terms of relations
between wet/dry variations and moisture budget in the
Pearl River basin (97º39′E–117º18′E; 3º41′N–29º15′N).
Figure 1 shows the location of the study region. Detailed
introduction of the Pearl River basin can be found in our
previous publications (e.g., Zhang et al. 2008b, 2009a).
Moisture flux analysis based on NCAR/NCEP dataset
indicated stronger intensity of water vapor transport in
rainy season than that in winter (dry season), showing
considerable influence of water vapor flux on dry and wet
conditions of the Pearl River basin (Zhang et al. 2008b).
And this finding was further corroborated by good
correlations identified between moisture budget, moisture
content and number of wet months in winter over the Pearl
River basin. Besides, Zhang et al. (2009a) also studied the
changes of precipitation concentration (CI) within the Pearl
River basin, indicating an increasing CI after 1990s when it
is compared to that before 1990s. We tentatively concluded
that these CI changes in the Pearl River basin are likely to
be associated with the consequences of the well-evidenced
global warming. There are a couple of studies addressing
the dry/wet conditions and precipitation concentrations in
the Pearl River basin. Moisture budget and water vapor flux
in wet and dry seasons were also analyzed with aim to
understand the background of large-scale circulation behind
changes of the wet/dry conditions in the Pearl River basin
(e.g. Zhang et al. 2008b). Seasonal variations of moisture
budget and water vapor flux, as one of the key components
in the hydrological cycle, are not thoroughly studied so far;
and this is not helpful for good understanding of hydrological cycle under the changing climate and the wellevidenced global warming in particular. In this case, we
attempt to analyze moisture budget and water vapor flux
based on the NCEP/NCAR reanalysis dataset with aim to
gain knowledge of changing properties of hydrological
cycle such as abrupt behaviors, trends and also the
periodicity characteristics. This will be of great scientific
and practical merits in understanding hydrological cycle
and also in sound river management under the changing
climate and intensifying human activities in the Pearl River
basin. Therefore, the major objectives of this study is to
understand statistical properties of moisture budget and
water vapor flux such as abrupt behaviors, trends, and also
and periodicity properties.
2 Data and methodology
2.1 Data
The whole layer moisture and associated transport properties are investigated by analyzing the NCAR/NCEP
reanalysis dataset covering 1948–2006. In the real atmo-
Water vapor budget in the Pearl River basin and its implications
187
Fig. 1 Location of the study
region, the Pearl River basin,
and the hydrological stations
and rain stations
sphere, the moisture is very low above 300 hPa, so that a
top of the atmosphere pressure of 300 hPa will be used in
the study (e.g., Miao et al. 2005; Zhang et al. 2008c). The
zonal moisture transport flux (QU), meridional moisture
transport flux (QV), and whole layer moisture budget (QT)
at regional boundaries are calculated based on the following
equations:
Z
1 p
Qu ðx; y; t Þ ¼
qðx; y; p; t Þuðx; y; p; t Þdp
ð1Þ
g ps
Qv ðx; y; t Þ ¼
QW ¼
82
X
1
g
Z
p
qðx; y; p; t Þvðx; y; p; t Þdp
Qu ðl1 ; y; t Þ
QE ¼
81
QS ¼
l2
X
ð2Þ
ps
82
X
Qu ðl2 ; y; t Þ
ð3Þ
Qv ðx; 8 2 ; t Þ
ð4Þ
81
Qv ðx; 8 1 ; t Þ
QN ¼
l1
QT ¼ QW QE þ QS QN
l2
X
l1
ð5Þ
Where u and v are the zonal and meridional components
of the wind field respectively; q is the specific humidity; x,
y, p, and t denote longitude, latitude, height and time
respectively. u(x,y,p,t) denotes zonal wind component at the
location of (x,y), the height of p and the time of t. ps is
surface pressure; p is atmospheric top pressure; g is
acceleration of the gravity; QW, QE, QS, QN are the west,
east, south, and north regional boundaries of the Pearl River
basin respectively; and 81, 82, l1, l2 are the latitude and
longitude according to the regional boundaries (Zhou et al.
1998; Miao et al. 2005; Zhang et al. 2008c). The four
boundaries defined for the Pearl River basin are based on
the longitude and latitude of the study river basin. To show
influences of propagation changes of water vapor flux on
precipitation variations and even the ground surface water
resource, we also explore the relations between water vapor
flux, precipitation, and streamflow variations within the
Pearl River basin. The West River is the largest tributary
accounting for 77.8% of the total drainage area of the Pearl
River basin; the North River is the second largest tributary
accounting for 10.3% of the total drainage area of the Pearl
River basin. The total streamflow of the West and the North
Rivers accounts for more than 99.7% of the total streamflow of the Pearl River basin, largely representing the
hydrological processes of the Pearl River basin. Therefore,
we analyzed the total streamflow of the Sanshui and the
Makou stations to show the streamflow variations of the
Pearl River basin. As for the precipitation changes, we
extracted precipitation data from the 160 rain gauging
stations which have good quality and continuous data
records for the period 1951–2005 (Gemmer et al. 2004;
Zhang et al. 2009b). The data are from the National
Climatic Centre of the China Meteorological Administration. We extracted precipitation dataset of 12 stations in the
Pearl River basin from those of China. Location of the rain
stations and the hydrological stations can be found in
Fig. 1.
2.2 Methodology
Methods used in this study are continuous wavelet
transform (CWT), the simple two-phase linear regression
scheme and linear regressive technique. The CWT tech-
188
Q. Zhang et al.
nique (Torrence and Compo 1998; Grinsted et al. 2004)
aims to analyze localized variations of power within a time
series. The wavelet transform has also been successfully
applied to analyze the hydrologic effects of the construction
and operation of dam on hydrological processes based on
changes of periodicity properties (White et al. 2005; Zhang
et al. 2008c). More recently, we used this method in study
of annual maximum streamflow series of the Yangtze
River basin (Zhang et al. 2007). Before CWT analysis, the
normality of the data series is first tested by the
Kolmogorov–Smirnov test. The method first compares
the specified theoretical cumulative distribution function
(e.g., the normal distribution in this study) with the sample
cumulative density function based on observations, then
calculates the maximum deviation, D, of the two. If, for
the chosen significance level, the observed value of D is
greater than or equal to the critical tabulated value of the
Kolmogorov–Smirnov statistic, the hypothesis of normal
distribution is rejected. After this step, continuous wavelet
transform is performed on meteor-hydrological dataset.
The CWT (Torrence and Compo 1998) is introduced
simply here. It is assumed that xn is a time series with equal
time spacing δt and n=0…N−1. y o(η) is a wavelet function
depending on a dimensionless ‘time’ parameter η with zero
mean and localized in both frequency and time (Farge
1992; Torrence and Compo 1998). Morlet wavelet is used
in this study due to the fact that Morlet wavelet provides a
good balance between time and frequency. The Morlet
wavelet is formulated as:
y o ðhÞ ¼ p 1=4 eiwoh eh =2
2
ð6Þ
where ωo is the non-dimensional frequency, here taken to
be 6 to satisfy the admissibility condition (Farge 1992;
Torrence and Compo 1998). The continuous wavelet
transform of xn is defined as the convolution of xn with a
scaled and translated version of y o(η):
;
N 1
X
ðn nÞdt
Wn ðsÞ ¼
xn ; y »
ð7Þ
s
n;
where the (*) indicates the complex conjugate. Because the
wavelet is not completely localized in time, to ignore the
edge effects the cone of influence (COI) was introduced.
Here, COI is the region of the wavelet spectrum in which
edge effects become important and is defined here as the efolding time for the autocorrelation of wavelet power at
each scale. This e-folding time is chosen so that the wavelet
power for a discontinuity at the edge drops by a factor e−2
and ensures that the edge effects are negligible beyond this
point (Grinsted et al. 2004; Torrence and Compo 1998).
The statistical significance of wavelet power can be
assessed under the null hypothesis that the signal is
generated by a stationary process being given the back-
ground power spectrum (Pk). It is assumed that the time
series has a mean power spectrum, possibly given by (8); if
a peak in the wavelet power spectrum is significantly above
this background spectrum, then it can be assumed to be a
true feature with a certain confidence. The “95% confidence interval” refers to the range of confidence about a
given value. To determine the 95% confidence level
(significant at 5%), one multiplies the background spectrum
(8) by the 95th percentile value for χ22 (Torrence and
Compo 1998). Many geophysical series have the red noise
characteristics which can be modeled by a first order
autoregressive AR(1) process. The Fourier power spectrum
of an AR(1) process with lag-1 autocorrelation α (e.g.,
Allen and Smith 1996) is given by (Grinsted et al. 2004) as
Pk ¼
1 a2
ð8Þ
2
j1 ae2ipk j
where k is the Fourier frequency index. Torrence and
Compo (1998) used the Monte Carlo method to show that
the probability that the wavelet power of a process with a
given power spectrum (Pk) is greater than p is
!
X 2
W ðsÞ
1
n
< p ¼ pk # 2v ðpÞ
ð9Þ
D
2
2
sX
where v is equal to 1 for real and 2 for complex wavelets.
The simple two-phase linear regression scheme was
proposed and used by Solow (1987), Easterling and
Peterson (1995), and Vincent (1998). In this study, we
introduced and modified this method based on the work by
Lund and Reeves (2002) so that it can help to reveal the
abrupt behaviors of the meteor-hydrological series in the
time scale vs. time and space.
The model was written as:
m1 þ a 1 t1 þ "t
Xt ¼
ð10Þ
m2 þ a 2 t2 þ "t
w h e r e t1 ¼ ½j n; j 1; t2 ¼ ½j; j þ n 1. T h e s u b sample size n is defined as n = 2, 3,…, <N/2. The quantity
j ¼ n þ 1; n þ 2; :::; N n þ 1 is the reference time point.
N is the length of the time series. The least squares
estimates of the trend parameters in Eq. 6 are obtained by:
ðt t 1 Þ X t X 1
j1
P
b1 ¼
a
t¼jn
and
j1
P
ðt t 1 Þ
2
t¼jn
jþn1
P
b2 ¼
a
ðt t 2 Þ Xt X 2
t¼j
jþn1
P
t¼j
ðt t 2 Þ2
ð11Þ
Water vapor budget in the Pearl River basin and its implications
In (7), X 1 and X 2 s are the average of the sub-series
before and after time j, respectively. t 1 and t 2 are the
average time observations before and after time j, respectively. Least squares estimates of the location parameters μ1
and μ2 in Eq. 6 are:
b1 t 1 and m
b2 t 2
b2 ¼ X 2 a
b1 ¼ X 1 a
m
ð12Þ
The denominators in (7) can be explicitly evaluated
as:
j1
X
ðt t 1 Þ2 ¼
t¼jn
¼
jþn1
X
ðj 1Þjðj 2Þ
and
ðt t 2 Þ2
12
t¼j
ð n j þ 1Þ ð n j þ 2Þ ð n j Þ
12
ð13Þ
Under the null hypothesis of no change points, the
regression parameters of the two phases must agree, i.e.,
b1 a
b2 should be
b1 m
b2 and a
α1 =α1 and μ1 =μ2. If so, m
close to zero for each sub-samples divided by j.
Rescaling this to a regression F statistic merely states
that (Lund and Reeves 2002)
Fc ¼
ðSSERed SSEFull Þðn 4Þ
2SSEFull
ð14Þ
In (10), SSEFull is the ‘full model’ sum of squared errors
computed from
SSEFull ¼
j1
X
b1 t Þ2 þ
b1 a
ðXt m
jþn1
X
t¼jn
b 2 t Þ2
b2 a
ðX t m
t¼j
ð15Þ
SSERed is the ‘reduced model’ sum of squared errors,
which was formulated as
SSERed ¼
jþn1
X
bRed t Þ2
bRed a
ðXt m
ð16Þ
t¼jn
bRed are estimated under the constraints
bRed and a
where m
bRed and m1 ¼ m2 ¼ m
bRed (Lund and Reeves
a1 ¼ a 2 ¼ a
2002). If a change point is present at time j−1, Fc should be
statistically large when compared to the threshold value by
F test. The effective degree of freedom after the correction
of dependence and in a normalized distribution for the time
series (Storch and Zwiers 1999; Jiang et al. 2007) can be
estimated by
2n
Ef f D ¼
INT 1 þ 2
INTP
ðn=2Þ
!
ð17Þ
rX ðt Þrt ðt Þ
t¼1
where INT denotes taking the integer part of the number.
After the effective degree of freedom is known, the
threshold value (Fth) can be obtained via the F test table
189
(Lund and Reeves 2002). If Fc >Fth, then we can say that
the change point is statistically present.
3 Results and discussions
Precipitation variations in the Pearl River basin are closely
associated with the East Asian summer monsoon. The west
Pacific (120°E–133°E) and the Bengal Bay are the major
sources of atmospheric moisture entering the Pearl River
basin and so does the Yangtze River basin (Zhang et al.
2008d), and this can be observed in Fig. 2. Figure 2 also
shows that most of the atmospheric moisture is coming
from the south and west edges of the Pearl River basin,
while the leaving atmospheric moisture largely exits
through the east and north edges. We analyzed the changing
properties of the water vapor flux entering into or departing
from the south, east, west, and north edges and also the net
moisture budget and net water vapor flux within the Pearl
River basin. Figure 3 shows the continuous wavelet
transform of water vapor flux series of the east
(Fig. 3(a1)), the west (Fig. 3(b1)), the south (Fig. 3(c1))
and the north (Fig. 3(d1)) boundaries of the Pearl River
basin. Generally, the propagation of the water vapor is
distinctly characterized by annual periods, more moisture
vapor flux in summer and less in winter, and this point can
be well supported by Fig. 2. Therefore, we could find
significant annual period. The wavelet power spectra for the
water vapor flux of the south, east and west boundaries
clearly show significant annual period. The power distribute consistently in the 1-year band, except the power for the
water vapor flux of the west boundary which distributes
intermittently during 1960–1970. Different changing properties can be found in the wavelet power spectra for the
water vapor flux via the north edge of the Pearl River basin
(Fig. 3(d2)). After the early 1980s, the annual periods
appear sporadically and intermittently. After the end of
1990s, the annual periods disappear completely. In comparison with the wavelet power spectra of the water vapor
flux of the other edges of the Pearl River basin, more
regions characterized by high wavelet power can be
observed in the <0.5-year bands and it is particularly true
after the 1980s. It is an interesting finding, which means
that the northward propagation of the water vapor flux via
the north edge of the Pearl River basin comes to be of
higher frequency after the early 1980s. Figure 3d1 also
indicates that the propagation of water vapor flux after the
early 1980s is characterized by smaller magnitude and
higher frequency.
Seasonal variations of water vapor flux of the four edges
are demonstrated in Fig. 4. Two change points are detected
in the water vapor flux changes in spring in the east edge
(Fig. 4(1)a). Before the early 1980s, the water vapor flux in
190
Q. Zhang et al.
(A)
(B)
(C)
(D)
(E)
Fig. 2 Water vapor flux (kilograms per meter per second) variations in a spring, b summer, c autumn, d winter, and e annual
spring is dominated by increase, and decreasing water
vapor flux can be observed after early 1980s. General
decrease can be found in the water vapor flux changes in
summer. Increasing water vapor flux in autumn and winter
can be found before the early 1960s (Fig. 4(1)b). No
obvious changes can be detected after the early 1960s.
Figure 4(2) illustrates changes of water vapor flux of the
west edge, showing somewhat complicated changing
properties when compared to those of the east edge. It can
be seen from Fig. 4(2) that 1960s can be seen as the turning
point from increase to decrease in three seasons. No visible
changes are detected within the changes of the water vapor
flux in spring after the 1960s. After the 1960s, the water
vapor flux in summer, autumn, and winter is dominated by
decreasing trends. It can be seen from Fig. 4(3) that the
water vapor flux of spring, summer, and autumn is
−2
1970
1980
1990
2000
Period (years)
1960
1960
1970
1980
Time (year)
1990
2000
2
0
−2 C1
−4
1950
0.25
0.5
1
2
4
8 C2
16
1950
1960
1970
1980
1990
2000
1960
1970
1980
Time (year)
1990
2000
2 B1
0
−2
−4
1950
1960
1970
1980
1990
2000
0.25
0.5
1
2
4
8 B2
16
1950
1960
1970
1980
Time (year)
1990
2000
Water vapor flux
(kg×m−1×s−1)
−4
1950
Water vapor flux
(kg×m−1×s−1)
0
191
Period (years)
Period (years)
2 A1
0.25
0.5
1
2
4
8 A2
16
1950
Water vapor flux
(kg×m−1×s−1)
Period (years)
Water vapor flux
(kg×m−1×s−1)
Water vapor budget in the Pearl River basin and its implications
4
D1
2
0
−2
1950
1960
1970
1980
1990
2000
0.25
0.5
1
2
4
8 D2
16
1950
1960
1970
1980
Time (year)
1990
2000
Fig. 3 Wavelet transform of water vapor flux variations traveling through the four boundaries defined in terms of the Pearl River basin in this study.
a East; b west; c south; d north. The U-shaped line shows cone of influence. The thick solid lines denote 95% confidence level using red noise model
characterized mainly by decreasing trends. However, water
vapor flux in winter is in increasing trends. We will discuss
the implications of this increasing water vapor flux in
winter via the south edge. Similar changing properties are
observed in the variations of the water vapor flux via the
north edge (Fig. 4(4)) except that the summer water vapor
flux is increasing before earlier 1960s. Another different
feature is that, as shown in Fig. 4(4)b, the increasing winter
vapor flux breaks down in middle 1970s and no visible
changes are found after the middle 1970s.
In this study, we also compute the areal net water vapor
flux with aim to understand its abrupt behaviors and trends
(Fig. 5). It can be found in Fig. 5 that the areal net water
vapor flux of spring, summer, and autumn is increasing
before early 1970s, and is decreasing after early 1970s. Two
change points were found within the changes of the areal
net water vapor flux of spring. After the second change
point, i.e., late 1980s, the areal net water vapor flux of
spring is increasing. Figure 5b indicates distinctly different
changing properties in terms of areal net water vapor flux in
winter as compared with other seasons. Two change points
are detected within the changes of the areal net water vapor
flux in winter (Fig. 5b), one occurred in 1979 and another
in 1992. Before 1979, it is decreasing and after 1979 the
areal net water vapor flux in winter is dominated by
increasing trends. Figure 5a demonstrates significant
decreasing areal net water vapor flux in summer after the
early 1970s. These changes of water vapor flux heavily
influenced the wet and dry conditions of the Pearl River
basin, causing drying tendency in rainy seasons and wetting
tendency in dry seasons. Figure 6 illustrates wavelet power
spectra of areal moisture budget of the Pearl River basin.
Significant wavelet power is observed during the mid1950s to approximately the early 1960s and mid-1960s to
approximately the late 1970s. After 1980, only sporadic
and intermittent distribution of high wavelet power can be
identified, showing considerable decrease of moisture
budget after 1980s. Besides, more regions of significant
wavelet power spectra appear in <0.5-year bands, showing
highly frequent variations of moisture budget after 1980s.
Figure 7 shows that the moisture budget of spring, summer,
and autumn is of similar changing properties, increasing
trends before early 1970s and decreasing trends after early
1970s. General decreasing trends are observed in the
changes of moisture budget in winter. Visual inspection of
changing curves of moisture budget in winter indicates
increasing moisture budget after the end of 1990s. All these
evidences tend to drive the Pearl River basin to be wetter in
winter and dryer in summer. Therefore, we can say that the
influences of climate change on hydrological cycle in the
Pearl River basin are the seasonal shifts of precipitation
variations.
Water vapor flux (kg×m−1×s−1)
80
A
60
40
Spring
20
0
Summer
1950
40
1960
1970
1980
Time (years)
1990
2000
B
20
Winter
0
−20
−40
Autumn
−60
1950
1960
1970
1980
Time (years)
1990
2000
Water vapor flux (kg×m−1×s−1) Water vapor flux (kg×m−1×s−1)
Q. Zhang et al.
Water vapor flux (kg×m−1×s−1)
192
100
A
Spring
50
0
Summer
1950
60
1960
A
50
40
30
Spring
1970
1980
Time (years)
20
1990
2000
B
Winter
10
0
−10
−20
Autumn
1950
1960
1970
1980
1990
2000
Time (years)
(3)
Water vapor flux (kg×m−1×s−1) Water vapor flux (kg×m−1×s−1)
Water vapor flux (kg×m−1×s−1)
Water vapor flux (kg×m−1×s−1)
Summer
1960
2000
1990
2000
Winter
B
20
0
Autumn
−20
1950
1960
1970
1980
Time (years)
(2)
60
1950
1990
40
(1)
70
20
1970
1980
Time (years)
60
A
Summer
50
40
30
20
10
Spring
1950
30
1960
1970
1980
Time (years)
1990
2000
B
Autumn
20
10
0
Winter
1950
1960
1970
1980
Time (years)
1990
2000
(4)
Fig. 4 Abrupt and trend behaviors of the water vapor flux traveling through (1): the east; (2) the west; (3) the south and (4): the north boundaries
of the Pearl River basin
To demonstrate impacts of propagation of water vapor
flux on the precipitation changes and streamflow, we
attempt to address relations between these three hydrological components within the Pearl River basin. Figure 8
shows that precipitation in spring, summer and autumn is
decreasing after approximately 1970s. Generally, two
turning points can be observed in the areal summer average
precipitation: 1960s and mid-1990s. It can be seen from
Fig. 8 that the areal average summer precipitation of the
Pearl River basin is dominated by decreasing trends after
about 1960s, which is in good agreement with the water
vapor flux changes as discussed above. Slight increase can
be observed in the changes of the winter precipitation.
Figure 9 demonstrates intuitively the relations between
precipitation and water vapor flux. It can be observed
obviously from Fig. 9 that the general tendency of
precipitation changes is in good line with that of the water
vapor flux variations, showing considerable influences of
water vapor flux on precipitation variations over the Pearl
River basin, which is in good line with the results we
obtained by the study on the wet/dry conditions of the Pearl
River basin and their associations with water vapor flux and
moisture content (Zhang et al. 2008b). We also quantitatively evaluate the relations between these two hydrological
components. The correlation analysis indicates significant
correlation between precipitation and water vapor flux in
autumn and winter, and the relations are not significant in
spring and summer. This might be due to the fact that more
than one factor influences the precipitation changes in
summer and spring, e.g., more typhoon-induced precipitation and convective precipitation in summer than other
seasons. We also analyzed relations between precipitation
and streamflow variations with the aim to indirectly
illustrate the influences of water vapor flux on streamflow.
4
2
0
Spring
A
−2
−4
Water vapor flux
(kg×m−1×s−1)
Summer
6
1950
1960
1970
1980
Time (years)
1990
2000
Autumn
10
5
B
0
1950
Winter
1960
1970
1980
Time (years)
1990
2000
Fig. 5 Abrupt and trend behaviors of the areal net water vapor flux
within the Pearl River basin. The denotations in this figure have the
same meaning as those of the above figures
Good relations are found between precipitation and streamflow variations (Fig. 10). The tendency of precipitation
changes matches well those of the streamflow variations.
Correlation analysis also indicates significant correlations
between precipitation and streamflow at >95% confidence
level. Based on what was mentioned above, we show
altered hydrological cycle within the Pearl River basin
reflected mainly by altered water vapor flux and moisture
budget. Decreasing northward propagation of water vapor
flux may heavily influence the precipitation and streamflow
across the Pearl River basin, leading to decreasing
precipitation and streamflow. Seasonal shifts of water vapor
flux, precipitation, and streamflow may arouse new
challenges in terms of river management in the Pearl River
basin.
Water vapor flux (g×m−2×s−1)
Water vapor flux
(kg×m−1×s−1)
8
193
Water vapor flux (g×m−2×s−1)
Water vapor budget in the Pearl River basin and its implications
30
A
Summer
20
10
0
Spring
−10
1950
40
1960
1970
1980
Time (years)
B
1990
2000
1990
2000
Autumn
30
20
Winter
10
0
1950
1960
1970
1980
Time (years)
Fig. 7 Abrupt and trend behaviors of the areal moisture budget within
the Pearl River basin
4 Conclusions
We thoroughly analyzed water vapor flux and moisture
budget based on the NCAR/NCEP reanalysis dataset with
aim to understand changing properties of these key
hydrological components by using the continuous wavelet
transform and the simple two-phase linear regression
scheme. We also attempted to address possible implications
of these changes in the hydrological cycle for the river
basin management within the Pearl River basin. We
analyzed the general linear trends of the time series
1.5
0.5
4
2
0
−2
1950
0.25
0.5
1
2
4
8
16
1950
1960
1970
1980
1990
2000
Standardized precipitation (mm)
Period (years)
Water vapor flux
(kg×m−1×s−1)
2
1
0
0.5
Spring
−0.5
1959 1971 1983 1995 2005
0.2
0
0
1959
Summer
1971
1983
1995 2005
0
Autumn
Winter
−0.2
−0.5
−0.4
1960
1970
1980
Time (year)
1990
2000
Fig. 6 Wavelet transform of the moisture budget in grams per square
meter per second in the Pearl River basin. The U-shaped line shows
cone of influence. The thick solid lines denote 95% confidence level
using red noise model
−0.6
−0.8
1959 1971 1983 1995 2005
−1
1959
1971
1983
1995 2005
Fig. 8 Seasonal changes of areal precipitation within the Pearl River
basin. The arrows in the figure show the trends of specific time
intervals
194
Q. Zhang et al.
1
10
2
0.5
5
1
5
0
0
0
0
−5
−1
A
−0.5
1960 1970 1980 1990 2000
−1
1960 1970 1980 1990 2000
−1
Water vapor flux (kg×m ×s )
0.5
C
0
10
B
−5
Precipitation (mm)
15
10
0
D
10
−0.5
−0.5
5
−1
0
0
1960 1970 1980 1990 2000
−1
1960 1970 1980 1990 2000
Fig. 9 Relations between seasonal water vapor flux and areal
precipitation in a spring; b summer; c autumn, and d winter
considered in this study if no change points are identified.
Some interesting conclusions are obtained as follows:
1. The water vapor is mainly from the South China Sea
and Bengal bay. Water vapor enters the Pearl River
basin mainly via the south and east edges. The water
vapor flux of the south and west edges are decreasing
and the turning points occurred in the early 1960s.
Significant decreasing trends are found in the changes
of water vapor flux propagating through the north edge
of the Pearl River basin. The wavelet transform spectra
also reveal that the monthly water vapor flux through
the north edge is decreasing reflected by intermittent
distribution of the wavelet power spectra after early
1980s. Besides, the periodicity properties of the water
vapor flux through the north edge imply that the
northward propagation of water vapor flux comes to
be highly frequent and be of smaller magnitude after
the early 1980s.
2. The results of this study indicate that altered hydrological cycle is mainly manifested by seasonal shifts of
water vapor flux after the early 1960s. These changes
of water vapor flux alter the seasonal variations of wet
and dry conditions across the Pearl River basin. Our
previous study indicated that the dry seasons come to
be wetter and rainy seasons (April–September) dryer
within the Pearl River basin (Zhang et al. 2008b). We
attribute the seasonal shifts of precipitation variations to
the seasonal shifts of water vapor flux. The areal net
water vapor flux is decreasing in the early 1970s. The
exception is the winter when the areal net water vapor
flux is increasing after the late 1970s. The moisture
budget is in similar changing properties. High wavelet
power is observed during 1952–1963 and 1963–1980.
Only sporadic and intermittent distribution of wavelet
power spectra are found after 1980s. Decreasing trends
of moisture budget are also identified after the early
1970s.
3. Correlation analysis between water vapor flux, precipitation, and streamflow indicates tremendous influences
of water vapor flux on precipitation variations and
streamflow. Significant correlations are observed between precipitation and water vapor flux in autumn and
winter, and not significant correlations in spring and
summer, implying more than one factor besides water
vapor flux on precipitation changes such as more
typhoon-induced and convective precipitation events
in summer than other seasons. Significant correlations
are found between streamflow and precipitation within
the Pearl River basin, indicating overwhelming impacts
of precipitation on ground surface water resource over
the Pearl River basin. In so doing, we also address
indirectly the impacts of water vapor flux on the water
resource. We can say that altered hydrological cycle
may trigger unexpected ecological problems by altering
hydrological processes and precipitations. The conclusions of this study will be of great scientific and
practical merits in basin scale water resource management in the Pearl River basin under the changing
climate, and global warming in particular. Besides, the
decreasing northward propagation of water vapor may
make the north China face good challenge in terms of
water resource management. Hydrological responses of
the river basins in north China, particularly the Yangtze
River basin and the Yellow River basin, should be
studied with aim to provide good scientific basis for
river basin management.
1
4
A
0.6
2
0.2
0
−0.2
−2
B
Streamflow (m3/s)
5
2
0
0
Precipitation (mm)
−0.6
1959
1
1974
1989
−4
2005
−2
1959
0
C
2
0
0
1974
1989
−5
2005
5
4
D
2
−0.5
0
−1
1959
−2
1974
1989
2005
−1
1959
1974
1989
−2
2005
Fig. 10 Relations between seasonal streamflow and areal precipitation changes in a spring; b summer; c autumn and d winter
Water vapor budget in the Pearl River basin and its implications
Acknowledgments This work was financially supported by the ‘985
Project’ (Grant No. 37000-3171315) and by a grant from the Research
Grants Council of the Hong Kong Special Administrative Region,
China (Project No. CUHK405308). Cordial thanks will be extended to
reviewers and the editor-in-chief, Prof. Dr. Hartmut Grassl for their
invaluable comments which greatly helped to improve the quality of
this paper.
References
Allen M, Ingram WJ (2002) Constraints on future changes in climate
and the hydrologic cycle. Nature 419:224–232
Allen MR, Smith LA (1996) Monte Carlo SSA: Detecting irregular
oscillations in the presence of coloured noise. J Climatol 9:3373–
3404
Brutsaert W, Parlange MB (1998) Hydrologic cycle explains the
evaporation paradox. Nature 396(5):30
Chaplot V (2007) Water and soil resources response to rising levels of
atmospheric CO2 concentration and to changes in precipitation
and air temperature. J Hydrol 337(1–2):159–171
Easterling DR, Peterson TC (1995) A new method for detecting
undocumented discontinuities in climatological time series. Int J
Climatol 15:369–377
Farge M (1992) Wavelet transform and their application to turbulence.
Annu Rev Fluid Mech 24:395–457
Gao G, Chen DL, Xu C-Y, Simelton E (2007) Trend of estimated
actual evapotranspiration over China during 1960-2002. J Geophys Res-Atmos 112:D11120. doi:10.1029/2006JD008010
Gemmer M, Becker S, Jiang T (2004) Observed monthly precipitation
trends in China 1951–2002. Theor Appl Climatol 77:39–45
Goyal RK (2004) Sensitivity of evapotranspiration to global warming:
a case study of arid zone of Rajasthan (India). Agr Water Manag
69:1–11
Grinsted A, Moore JC, Jevrejeva S (2004) Application of the cross
wavelet transform and wavelet coherence to geophysical time
series. Nonlinear Process Geophys 11:561–566
Heideman KF, Fritsch JM (1988) Forcing mechanisms and other
characteristics of significant summertime precipitation. Weather
Forecasting 3:115–130
Jiang JM, Gu XQ, Ju JH (2007) Significant changes in subseries
means and variances in an 8000-year precipitation reconstruction
from tree rings in the southwestern USA. Ann Geophys 25:1–12
Karl TR, Knight RW, Easterling DR, Quayle RG (1996) Indices of
climate change for the United States. Bull Am Meteorol Soc
77:279–292
Kruijt B, Witte J-PM, Jacobs CMJ, Kroon T (2008) Effects of rising
atmospheric CO2 on evapotranspiration and soil moisture: a
practical approach for the Netherlands. J Hydrol 349(3–4):257–267
Labat D, Goddéris Y, Probst JL, Guyot JL (2004) Evidence for global
runoff increase related to climate warming. Adv Water Resour
27:631–642
Lund R, Reeves J (2002) Detection of undocumented changepoints: a
revision of the two-phase regression model. J Climate 15:2547–2554
195
Menzel L, Bürger G (2002) Climate change scenarios and runoff
response in the Mulde catchment (Southern Elbe, Germany). J
Hydrol 267:53–64
Miao QJ, Xu XD, Zhang SY (2005) Whole layer water vapor budget
of Yangtze River valley and moisture flux components transform
in the key areas of the plateau. Acta Meteorological Sinica
63:93–99
Milly PCD, Dunne KA, Vecchia AV (2005) Global pattern of trends in
streamflow and water availability in a changing climate. Nature
438:347–350
Semenov V, Bengtsson L (2002) Secular trends in daily precipitation
characteristics: greenhouse gas simulation with a coupled
AOGCM. Climate Dynamics 19:123–140
Solow AR (1987) Testing for climate change: an application of the twophase regression model. J Clim Appl Meteorol 26:1401–1405
Storch HV, Zwiers F (1999) Statistical analysis in climate research.
Cambridge University Press, Cambridge, p 116
Thodsen H (2007) The influence of climate change on stream flow in
Danish rivers. J Hydrol 333(2–4):226–238
Torrence C, Compo GP (1998) A practical guide to wavelet analysis.
Bull Am Meteorol Soc 79:61–78
Vincent LA (1998) A technique for the identification of inhomogeneities in Canadian temperature series. J Clim 11:1094–1104
White AM, Schmidt CJ, Topping JD (2005) Application of wavelet
analysis for monitoring the hydrologic effects of dam operation:
Glen Canyon Dam and the Colorado River at Lees Ferry,
Arizona. River Res Appl 21:551–565
Zhang Q, Xu C-Y, Jiang T, Wu YJ (2007) Possible influence of ENSO
on annual maximum streamflow of Yangtze River, China. J
Hydrol 333:265–274
Zhang Q, Xu C-Y, Chen YD (2008a) Spatial and temporal variability
of precipitation maxima during 1960-2005 in the Yangtze River
basin and possible association with large-scale circulation. J
Hydrol 353:215–227
Zhang Q, Xu C-Y, Zhang Z (2008b) Observed changes of drought/
wetness episodes in the Pearl River basin, China, using the
standardized precipitation index and aridity index. Theoretical and
Applied Climatology doi:10.1007/s00704-008-0095-4 (in press)
Zhang Q, Chen G, Su B, Disse M, Jiang T, Xu C-Y (2008c)
Periodicity of sediment load and runoff in the Yangtze River
basin and possible impacts from climatic changes and human
activities. Hydrol Sci J 53(2):457–465
Zhang Z, Zhang Q, Xu C-Y, Liu C-L and Jiang T (2008d)
Atmospheric moisture budget and floods in the Yangtze River
basin, China. Theoretical and Applied Climatology doi:10.1007/
s00704-008-0010-z (in press)
Zhang Q, Xu C-Y, Gemmer M, Chen YD, Liu C-L (2009a) Changing
properties of precipitation concentration in the Pearl River basin,
China. Stoch Environ Res Risk Assess 23:377–385
Zhang Q, Xu C-Y, Zhang Z, Chen YD, Liu C-L (2009b) Spatial and
temporal variability of precipitation over China, 1951–2005.
Theor Appl Climatol 95(1–2):53–68
Zhou J, Xue YF, Liu XF (1998) The source/sink distribution of water
vapor with its transfer in Asian monsoon region in August, 1994.
J Trop Meteorol 14(1):91–96
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement