Multitemporal Landsat Image Based
Water Quality Analyses of Danjiangkou Reservoir
Yinuo Zhang, Xin Huang, Wei Yin, and Dun Zhu
Abstract
Danjiangkou Reservoir (
DJKR
) is one of the largest artificial
freshwater lakes in Asia and a water source of the South:
the North Water Transfer Project. However, few studies have
analyzed the spatio-temporal water quality distribution or
investigated the causative factors of the long-term water qual-
ity variation of
DJKR
. In this study, we used multi-temporal
Landsat images combined with the multiple linear stepwise
regression (
MLSR
) method to retrieve long-term distributions of
the main water quality parameters in
DJKR
, i.e., total nitrogen
(
TN
), total phosphorus (
TP
), permanganate index (
COD
Mn
),
and five-day biochemical oxygen demand (
BOD
5
). Results
indicated the heavily polluted regions and an alarming water
quality deterioration trend between May 2006 and May 2014.
A combination of land use/land cover (
LULC
) maps and socio-
economic data was considered to investigate the causative
factors of the water quality distribution, as well as the deterio-
ration. This study could provide a valuable reference for the
decision-making for water quality conservation in
DJKR
.
Introduction
Freshwater is indispensable for our lives and daily activities.
However, China comprises 22 percent of the World’s popula-
tion but contains only 7 percent of the total surface freshwater
on Earth (Li
et al
., 2009). Influenced by the monsoon climate
and the mismanagement of water and soil resources, the water
distribution in China is highly heterogeneous. In other words,
less than 20 percent of the freshwater distributes in North
China, which accounts for 63.5 percent of China’s land area.
As a consequence, the North China Plain contains 0.35 billion
people, yet has per capita water resources of only 456 m
3
,
which is less than one-quarter of China’s average. Therefore,
the South-North Water Transfer (
SNWT
) Project was officially
launched in 2002 to solve this problem. The project has been
one of the largest strategic projects in China since 1949 and
has received global attention. This formidable and arduous
project has three routes: the eastern and middle routes aim to
channel water to North China, and the western route diverts
water to Northwest China. Danjiangkou Reservoir (
DJKR
),
located on the Han River, which is the longest tributary of the
Yangtze River, is the one of the largest artificial freshwater
lakes in Asia, with a surface area of ~1000 km
2
and a volume
of ~29 billion m
3
. It was therefore chosen to be the water
source of the middle route of the
SNWT
Project, which aims
to supply up to 13.8 billion m
3
of freshwater annually to the
North China Plain, including two municipalities (Beijing and
Tianjin), and more than 130 other cities, for domestic, indus-
trial, and agricultural use (Li and Zhang, 2005).
In addition,
DJKR
is one of the water sources of “NongFu
Spring”, which has been one of the most popular drinking wa-
ter brands of China since 1996 and produces over 0.6 million
tons of natural drinking water annually. The water quality of
DJKR
directly affects the drinking water security of hundreds
of millions of Chinese people and the implementation of the
largest-ever water transfer project. Therefore, periodic and ef-
ficient water quality monitoring in
DJKR
is urgently needed.
Traditional
in-situ
measurements are able to provide
details of the optical properties of water, and they provide
accurate data at fixed sample sites in
DJKR
. Nevertheless, this
approach is not only costly and time-consuming, but also
restricted by natural conditions, e.g. weather and terrain
(Guan
et al
., 2011). Moreover, traditional
in-situ
measure-
ments cannot provide the spatio-temporal distributions of the
water quality parameters (Chen
et al
., 2015), and hence limit
the comprehensiveness of the water quality monitoring. With
the advent of satellite images, they have been widely used for
inland water quality monitoring due to their extraordinary
ability of providing a synoptic view of water properties over
a large-scale spatial area (Chen and Quan, 2012). Landsat
imagery are generally applied in this situation, as they feature
a global coverage, the longest record of Earth observation, free
access, high-resolution, as well as multispectral data (Love-
land and Dwyer, 2012). The instructive application of multi-
temporal Landsat images in previous studies has confirmed
their potential in large-scale water quality monitoring. For
instance, Lathrop and Richard (1992), Kloiber
et al
. (2002),
Ritchie (2003), and McCullough
et al
. (2012) used multi-tem-
poral Landsat images to perform long-term analyses of water
clarity. Pastorguzman
et al
. (2015) and Tebbs
et al
. (2013)
applied Landsat
ETM+
bands to estimate chlorophyll-a (
Chl-a
)
concentration and successfully related the results to the local
algal blooms. Brezonik
et al
. (2005) made a characterization
of the optical properties between
Chl-a
and colored dissolved
organic matter (
CDOM
) using empirical models. Recently, Lobo
et al
. (2015) proposed a non-linear empirical regression model
to estimate
TSS
in the Tapajós River Basin, and then combined
it with the impact of gold mining activities.
However, little attention has been paid to the application
of satellite images in
DJKR
. In addition, the existing pertinent
studies have provided an insight mainly into
Chl-a
,
CDOM
,
and water clarity, but they have neglected the many other
important water quality parameters, such as total nitrogen
(
TN
), total phosphorus (
TP
), permanganate index (
COD
Mn
), and
five-day biochemical oxygen demand (
BOD
5
), which are also
closely related to anthropogenic activities and contribute to
the eutrophication of the lakes and reservoirs.
The purpose of this study was to apply multi-temporal and
multi-sensor Landsat images (
TM
,
ETM+
, and
OLI
) of
DJKR
from
Yinuo Zhang and Xin Huang are with the State Key
Laboratory of Information Engineering in Surveying, Mapping
and Remote Sensing, Wuhan University, Luoyu Road No.129,
Wuhan, Hubei Province, P.R.China (corresponding author:
Xin Huang,
.
Wei Yin and Dun Zhu are with the Yangtze River Water
Resources Protection Science Institute, Qintai Avenue No.
515, Wuhan, Hubei Province, P.R.China.
Photogrammetric Engineering & Remote Sensing
Vol. 83, No. 9, September 2017, pp. 643–652.
0099-1112/17/643–652
© 2017 American Society for Photogrammetry
and Remote Sensing
doi: 10.14358/PERS.83.9.643
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
September 2017
643
