Impact of Vegetation Change and Climate Variability on Runoff in
the Jingchuan Watershed in the Loess Plateau of China
Runqi Zhao, Mingfang Zhang
*
and Yiping Hou
School of Resources and Environment, University of Electronic Science and Technology of China, Chengdu, China 611731.
Keywords: Runoff, vegetation change, climate variability, the Loess Plateau
Abstract: Vegetation change and climate variability are often two main drivers for runoff variation. However, other
factors such as human activities (e.g., road construction, agriculture, dams, urbanization) may also yield
significant impact on runoff in populated watersheds. In this study, we selected the Jingchuan watershed, a
sub-watershed of the Jing River Basin in the Loess Plateau as an example, to quantify runoff variation
attributed to vegetation change, climate variability and other factors on runoff based on a single watershed
approach. Firstly, we applied modified double mass curve (MDMC) and Autoregressive Integrated Moving
Average (ARIMA) intervention to identify the impact of climate variability attributed to runoff variation. A
significant breakpoint was detected in 1998. Then, multivariate Autoregressive Integrated Moving Average
(ARIMAX) was used to differentiate the impact of vegetation change and other factors from non-climatic
factors. The results showed that the average annual runoff attributed to vegetation change, climate
variability and other factors were -16.61mm, 8.91mm and -1.33mm, respectively, resulting in a 9.03mm
reduction in annual runoff over the period of 1998-2003 comparing with the runoff without impacts. These
findings are beneficial to water supply and vegetation management in semi-arid areas such as Loess Plateau
of China.
1 INTRODUCTION
The impact of vegetation change on runoff has been
studied for over 100 years (Stednick, 1996;
Andréassian, 2004; Brown et al., 2005). Conclusions
from small watersheds (less than 100 km
2
) show that
vegetation change can significantly influence annual
runoff by altering evapotranspiration (Bosch and
Hewlett, 1982; Jones and Post, 2004). However, the
impact of vegetation change on hydrology in large
watersheds (more than 1000 km
2
) has some
inconsistent results, which shows both positive and
negative effect on runoff (Costa et al., 2003;
Siriwardena et al., 2006; Lin and Wei, 2008; Zhang
et al., 2012). The lack of an efficient, commonly-
accepted methodology can constrain forest
hydrological studies in large watersheds.
The Jingchuan watershed is located in the Loess
Plateau, with severe soil erosion and water scarcity.
Our objective of this study is to quantify runoff
variation attributed to climate variability, vegetation
change and other factors by use of a single
watershed approach, which can help us understand
the effect of vegetation recovery on runoff, and
control soil erosion in the Loess Plateau and
eventually support water resource management in
dry regions.
2 MATERIALS AND METHODS
2.1 Study Watershed and Data
The Jingchuan watershed, a sub-watershed of the
Jing River Basin, is located in Loess Plateau (Figure
1). It covers an area of 3155.12 km² and belongs to
temperate continental climate. The average annual
temperature is 9 °C, annual precipitation reaches to
525.5 mm, and majority of which falls in wet season
(May-October). The Jingchuan watershed is
dominantly covered by farmland and grassland,
which are 42.3% and 36.5% of total watershed area.
The Jing River Basin experienced afforestation since
late 1990s to prevent severe soil erosion.
Zhao, R., Zhang, M. and Hou, Y.
Impact of Vegetation Change and Climate Variability on Runoff in the Jingchuan Watershed in the Loess Plateau of China.
In Proceedings of the International Workshop on Environment and Geoscience (IWEG 2018), pages 529-532
ISBN: 978-989-758-342-1
Copyright © 2018 by SCITEPRESS – Science and Technology Publications, Lda. All r ights reserved
529

Figure 1: The location of the Jingchuan watershed.
Hydrological, climate and LAI data used in this
study dated back to 1983. The hydrological records
were obtained from the Jingchuan hydrological
station, and monthly runoff and precipitation data
were collected to calculate annual and seasonal data
series. According to historical records, the average
annual runoff only 39.5mm, and the hydrological
year can be divided into dry season (November-
April) and wet season (May-October). Meanwhile,
two active climate stations are available in the study
watershed, Kongtong station is located in the
upstream and Xifeng in the downstream. Monthly
mean/min/max temperature records in the two
stations were obtained from Climatic Data Center,
China Meteorological Administration (CMA). In
this study, LAI data from GLASS (Global Land
Surface Satellite Products) products were used to
indicate the vegetation variation characteristics.
2.2 Method
Climate variability, vegetation change and other
factors such as human activities are main factors for
runoff variation. Firstly, modified double mass
curve (MDMC) was used to separate climate
variability and non-climate factors (Wei and Zhang,
2010). In modified double mass curve (MDMC), the
y-axis is Qa which represents cumulative runoff and
the x-axis is Pae which represents cumulative
effective precipitation (the difference between
precipitation and actual evapotranspiration).
Seasonal data series were applied in this study. If
non-climatic factors have no significant effect on
runoff, the MDMC should be a straight line while a
significant breakpoint can be found if non-climate
factors made significant effect on runoff. Moreover,
we employed the Autoregressive Integrate Moving
Average (ARIMA) Intervention model to examine
the significance of the breakpoint. Then, predicted
line can be established using linear regression with
the data before the breakpoint. The difference
between observed line and predicted line refers to
runoff variation caused by non-climate factors
( ∆𝑄

). Finally, the multivariate time series
analysis (ARIMAX) is used to separate the
vegetation change and other factors from non-
climate factors (Hou et al., 2018). By taking the
cumulative LAI changes and ∆𝑄

as input, we can
get the predict ∆𝑄

. The difference between
∆𝑄

and ∆𝑄

expresses the influence of others
(other factors and statistical errors). The 95%
confident interval (CI) is adopted to remove the
statistical errors. In this way, runoff variation
attributed to other factors (∆𝑄

) and the vegetation
change (∆𝑄

) can be quantified.
3 RESULTS
A breakpoint was found in the year 1998 (Figure 2).
Additional, ARIMA Intervention model (Table 1)
confirmed the significance of the breakpoint. Using
linear regression, predicted line was calculated.
According to the calculation, seasonal runoff
variation caused by non-climate factors varied from
-79.25mm to 10.88mm, while, the climate
variability effect on seasonal runoff ranging from -
22.33mm to 90.97mm.
Figure 2: Modified double mass curve (MDMC) in the
Jingchuan watershed. In which, Qa is cumulative runoff,
Pae cumulative effective precipitation.
IWEG 2018 - International Workshop on Environment and Geoscience
530

Table 1: ARIMA Intervention model.
Model Input Model structure
Parameter estimation
p(1)
a
Ω(1)
b
Δ(1)
b
Slope of
MDMC
ARIMA Intervention:
ln(x), (1,0,0),
intervention at Year 1998
-0.33
(p=0.035)
-1.74
(p=0.000)
-1.00
(p=0.000)
a
The autoregressive parameter;
b
Parameters for intervention.
While seasonal runoff variation attributed to
climate variability were quantified by MDMC, we
used the ARIMAX model to separate the effect of
vegetation change and other factors from non-
climate factors. Table 2 shows the parameters of the
selected ARIMAX model. Figure 3 demonstrates the
differences between observed (∆𝑄

) and predicted
accumulated seasonal runoff variation for the non-
climatic factors (∆𝑄

), which is modelled by
ARIMAX model.
Moreover, we use 95% CI to eliminate statistical
errors. There were 3 data series within the 95%CI
(1999-wet, 2001-wet and 2002-dry), which
suggested that runoff deviation attributed to other
factors in these seasons can be ignored. After that,
runoff variation attributed to vegetation change and
other factors can be quantified.
Table 3 illustrates the quantification results from
1998 to 2003 in the Jingchuan watershed. Annual,
dry season and wet season runoff variation caused
by climate variability are 8.91mm, -16.40mm and
34.22mm, respectively, indicating that vegetation
change can significantly decrease annual and wet
season runoff (16.61mm and 37.22mm) and increase
dry seasonal runoff (4.01mm).
Table 2: ARIMAX model of seasonal runoff variation
attributed to non-climate factors.
Model Input
Parameter Estimation
p(1)
a
ΔLAIa
b
ΔQanc:
In ARIMA(1,0,0)
+ ΔLAIa
0.832
(p=0.0004)
-168.059
(p<0.0001)
a
The autoregressive parameter;
b
The accumulated
LAI changes.
Figure 3: Observed and predicted cumulative runoff
variation for the non-climatic factors.
Table 3: The quantification of different factors attributing to runoff variation.
Season
𝛥𝑄

(mm)
a
𝛥𝑄

(mm)
b
𝛥𝑄

(mm)
c
𝛥𝑄(mm)
d
LAI Reference LAI
DRY 1.55 4.01 -16.40 -10.83 0.26 0.26
WET -4.22 -37.22 34.22 -7.22 0.98 0.92
ANNUA
L
-1.33 -16.61 8.91 -9.03 0.62 0.59
a
Runoff variation attributed to other factors;
b
Runoff variation attributed to vegetation change;
c
Runoff variation attributed to climate variability;
d
The total runoff variation.
Impact of Vegetation Change and Climate Variability on Runoff in the Jingchuan Watershed in the Loess Plateau of China
531

4 DISCUSSION
Vegetation recovery by Grain for Green program in
the Jingchuan watershed yielded a negative effect on
annual runoff, a 5% increment in LAI averagely
resulted in about 42% reduction in annual runoff. A
similar study in the Loess Plateau also showed that
land use/cover change (reforestation) produced over
50% of the reduction in annual runoff ( Zhang et al.,
2008). Actually, transpiration of planted trees and
grass in the Loess Plateau are relative high, leading
to the reduction of annual runoff (Li et al., 2016;
Jian et al., 2015; Duan L et al., 2016). Therefore, in
the process of ecological restoration in Loess
Plateau, we should choose vegetation type with low
transpiration and less water consumption to reduce
the negative impact on water supply.
5 CONCLUSIONS
Annual and seasonal runoff in the Jingchuan
watershed are sensitive to vegetation change, a slight
increase in LAI due to forest and grass planting
leading to high transpiration. Meanwhile, climate
variability and vegetation change yield offsetting
effects suggesting a relative stable water supply in
the study area. Our results also demonstrate a great
implication for water resource management and
ecological restoration in the Loess Plateau.
ACKNOWLEDGMENT
This research was supported by National Key
Research and Development Program of China (No.
2017YFC0505006) and China National Science
Foundation (No. 31770759).
REFERENCES
Andréassian V 2004 Waters and forests: from historical
controversy to scientific debate Journal of Hydrology
291 1
Bosch J M and Hewlett J D 1982 A review of catchment
experiments to determine the effect of vegetation
changes on water yield and evapotranspiration J
Hydrology 55 1
Brown A E, Zhang L, Mcmahon T A, Western A W and
Vertessy R A 2005 A review of paired catchment
studies for determining changes in water yield
resulting from alterations in vegetation Journal of
Hydrology 310 1
Costa M H, Botta A and Cardille J A 2003 Effects of
large-scale changes in land cover on the discharge of
the Tocantins River, Southeastern Amazonia Journal
of Hydrology 283 1
Duan L, Huang M and Zhang L 2016 Differences in
hydrological responses for different vegetation types
on a steep slope on the Loess Plateau China Journal of
Hydrology 537(537)
Hou Y, Zhang M, Meng Z, Liu S, Sun P, Yang T 2018
Assessing the Impact of Forest Change and Climate
Variability on Dry Season Runoff by an Improved
Single Watershed Approach: A Comparative Study in
Two Large Watersheds, China Forests 9 1
Jian S, Zhao C, Fang S and Yu K 2015 Effects of different
vegetation restoration on soil water storage and water
balance in the Chinese Loess Plateau Agricultural &
Forest Meteorology 206
Jones J A and Post D A 2004 Seasonal and successional
streamflow response to forest cutting and regrowth in
the northwest and eastern United States Water
Resources Research 40 5
Li S, Liang W, Fu B, Lü Y, Fu S and Wang S 2016
Vegetation changes in recent large-scale ecological
restoration projects and subsequent impact on water
resources in China's Loess Plateau Science of the Total
Environment s 569–570
Lin Y and Wei X 2008 The impact of large-scale forest
harvesting on hydrology in the Willow watershed of
Central British Columbia Journal of Hydrology 359 1–
2
Siriwardena L, Finlayson B L and Mcmahon T A 2006
The impact of land use change on catchment hydrology
in large catchments: The Comet River, Central
Queensland, Australia Journal of Hydrology 326 1
Stednick J D 1996 Monitoring the effects of timber
harvest on annual water yield Journal of Hydrology
176 s 1–4
Wei X and Zhang M 2010 Quantifying streamflow change
caused by forest disturbance at a large spatial scale: A
single watershed study Water Resources Research 46
12
Zhang M, Wei X, Sun P and Liu S 2012 The effect of
forest harvesting and climatic variability on runoff in a
large watershed: The case study in the Upper Minjiang
River of Yangtze River basin Journal of Hydrology s
13
464–465
Zhang X, Zhang L, Zhao J, Rustomji P and Hairsine P
2008 Responses of streamflow to changes in climate
and land use/cover in the Loess Plateau, China Water
Resources Research 44 7
IWEG 2018 - International Workshop on Environment and Geoscience
532