Water Research 260 (2024) 121946

Contents lists available at ScienceDirect

Water Research
journal homepage: www.elsevier.com/locate/watres

Effects of landscape changes on water quality: A global meta-analysis

Xinying Shi a , Dehua Mao a, * , Kaishan Song a , Hengxing Xiang a , Sijia Li a , Zongming Wang a, b
a
State Key Laboratory of Black Soils Conservation and Utilization, Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun 130102,
China
b
National Earth System Science Data Center, Beijing 100101, China

A R T I C L E I N F O A B S T R A C T

Keywords: Landscape changes resulting from anthropogenic activities and climate changes severely impact surface water
Landscape changes quality. A global perspective on understanding their relationship is a prerequisite for pursuing equity in water
Water quality security and sustainable development. A sequent meta-analysis synthesizing 625 regional studies from 63
Meta-analysis
countries worldwide was conducted to analyze the impacts on water quality from changing landscape compo­
Water security
Global scale
sitions in the catchment and explore the moderating factors and temporal evolution. Results exhibit that total
nitrogen (TN), total phosphorus (TP), and chemical oxygen demand (COD) in water are mostly concerned and
highly responsive to landscape changes. Expansion of urban lands fundamentally degraded worldwide water
quality over the past 20 years, of which the arid areas tended to suffer more harsh deterioration. Increasing forest
cover, particularly low-latitude forests, significantly decreased the risk of water pollution, especially biological
and heavy metal contamination, suggesting the importance of forest restoration in global urbanization. The effect
size of agricultural land changes on water quality was spatially scale-dependent, decreasing and then increasing
with the buffer radius expanding. Wetland coverage positively correlated with organic matter in water typified
by COD, and the correlation coefficient peaked in the boreal areas (r=0.82, p<0.01). Overall, the global impacts
of landscape changes on water quality have been intensifying since the 1990s. Nevertheless, knowledge gaps still
exist in developing areas, especially in Africa and South America, where the water quality is sensitive to land­
scape changes and is expected to experience dramatic shifts in foreseeable future development. Our study
revealed the worldwide consistency and heterogeneity between regions, thus serving as a research roadmap to
address the quality-induced global water scarcity under landscape changes and to direct the management of land
and water.

1. Introduction substantial aggravation and unevenness resulting from regional diver­

gence in geographic and socioeconomic conditions (Sun et al., 2020;
The global landscape has undergone profound changes in recent Radwan et al., 2021). Continuous landscape changes have now sur­
decades due to the accelerating development of human societies, passed climate change as the primary factor influencing ecological
coupled with climate changes that, in turn, drive further alternations in processes in watersheds (Dale, 1997; Delpla & Rodriguez, 2014),
climate and disturb the provision of ecosystem services (Pielke Sr, 2005; significantly affecting regional water quality by impacting the biogeo­
Song et al., 2018; Olsson et al., 2019). Human-induced landscape chemical and hydrological cycles, posing challenges to interregional
changes, especially brought about by agricultural production and ur­ environmental equity and worldwide water security (Bullard et al.,
banization, have affected nearly three-quarters of the global terrestrial 1966; Peiffer et al., 2021).
area and the transformations continuously intensified (Winkler et al., Water security has been at the center of the world’s attention as a
2021; Zalles et al., 2021). Nevertheless, natural landscapes, particularly necessity for supporting lives and the well-being of humans (Vörösmarty
forests and wetlands, have faced pervasive loss over the past centuries et al., 2010). However, the handily available freshwater in ecosystems
(Mao et al., 2018a; Mao et al., 2018b; Fluet-Chouinard, 2023). Although encompassing rivers, lakes, and wetlands constitutes a fraction of less
land degradation has received concordant perception from international than 1 % of the total water volume and faces continuous endangerment
governments (Olsson et al., 2019), global landscape changes manifest (USGS, 2019). Saving the water quality is critical for alleviating water

* Corresponding author.
E-mail address: [email protected] (D. Mao).

https://doi.org/10.1016/j.watres.2024.121946
Received 21 February 2024; Received in revised form 12 June 2024; Accepted 13 June 2024
Available online 14 June 2024
0043-1354/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
X. Shi et al. Water Research 260 (2024) 121946

scarcity, salvaging Sustainable Development Goals, and also supporting the global context.
food and energy security (Wang et al., 2022a). Landscape changes have In this paper, a global meta-analysis is performed to address the
emerged as a major contributor to the deterioration of water quality in following questions: (1) What are the study trends and hotspots of the
the Anthropocene, marked by the reinforced contamination inputs from effects of landscape change on surface water quality until now? (2) What
land use activities and a weakened capacity of the natural landscape for are the overall correlations between kinds of landscape changes and
decontamination (Bullard, 1966; IFPRI & VEOLIA, 2015). Examples water quality? (3) What are the moderators in their correlations? And
include excess nutrients such as nitrogen and phosphorus in water how do the correlations vary under different moderators? This paper
bodies caused by agricultural inputs, and elevated bacteria concentra­ initially identifies the worldwide consistency and heterogeneity be­
tions discharged from urban sewage (Carpenter et al., 1998; Kang et al., tween regions regarding the relationship between landscape changes
2010). At the same time, the conversion from natural landscapes to and surface water quality, underpinning global knowledge and equity in
anthropogenic land covers leads to soil erosion, transporting sediments water security.
and organic carbon into receiving water and reducing pollutant purifi­
cation capacity (Sweeney et al., 2004; Tan et al., 2022). However, 2. Materials and methods
whether the impacts of worldwide uneven landscape changes on water
quality are consistent and whether they have become stronger in recent 2.1. Data collection and compilation
years remains unknown.
Copious regional-scaled studies have been conducted on the effects To collect data for analysis, over 20,000 peer-reviewed publications
of landscape changes on water quality with a wide spectrum of methods, were queried until December 2022 through Web of Science, Science­
indicating that landscape patterns broadly contribute to the formation, Direct and China National Knowledge Infrastructure Databases. Dupli­
release, interception and decomposition of pollutants in water (Bullard, cates and inaccessible articles were removed by machine screening,
1966; Simpson et al., 2022; Tan et al., 2022). In the majority of these while irrelevant articles were removed from the preliminary scanning of
studies, changes in the landscape were focused on land use/land cover titles and abstracts. Later the 625 selected studies were full-text scanned
(LULC) shifts (Mehaffey et al., 2005; Flood et al., 2022). The data source and compilated with their characteristic information (Fig. 1). A Detailed
of landscape changes experienced an evolution from rough field surveys flow diagram is shown in Appendix 1. Landscape compositions of in­
or historical planning documents (Hakamata et al., 1992; Broussard, terest were agricultural lands (AG), urban lands (UR), forests (FO),
2009), to broadscale, dynamic and high-resolution remote sensing im­ grasslands (GR), wetlands (WET), water bodies and barren lands.
ageries in recent years (Rimba et al., 2021; Zhang et al., 2022). The Herein, landscape changes refer to the change in proportion of each
surface water quality could be revealed by levels of various indicators composition in a certain watershed. Concerned water quality parame­
like pH, electric conductivity, concentrations of solids, nutrients, bac­ ters included acidity (pH), electrical conductivity (EC), dissolved oxygen
teria, heavy metals, etc. (Lee et al., 2009; Chiang et al., 2021; Flood (DO), total suspended solids (TSS), total dissolved solids (TDS), total
et al., 2022) or composite water quality indices (WQI) (Zhang et al., nitrogen (TN), nitrate, ammonia, total phosphorus (TP), phosphate,
2022). Commonly used quantitative methods to examine the response of chemical oxygen demand (COD), total biological oxygen demand (BOD),
water quality to landscape changes were correlation analysis (Chiang dissolved organic carbon (DOC), coliform, metal ions (K, Ca, Na, Mg,
et al., 2021), principal components analysis (Galbraith & Burns, 2007), etc.) and heavy metals (Cu, Zn, Mn, Pb, etc.).
redundancy analysis (Sliva & Williams, 2001), and linear regression From each of the studies, we extracted the correlation coefficients of
analysis (Mehaffey et al., 2005). As machine learning algorithms landscape composition changes versus water quality indicators as effect
developed, some studies have employed them concerning complex sizes in the meta-analysis, which emphasized the interactions between
environmental variables, represented by the random forest model variations (Borenstein et al., 2009), since the absolute changes of water
(Zhang et al., 2022; Xu et al., 2023) and the support vector machine quality parameters have tremendous discrepancy worldwide due to the
model (Luo et al., 2020). inherent geographical differences in the global water environments. A
However, there is still a gap in research globally addressing the set of 3228 data from 179 studies was finally extracted according to the
consistency and heterogeneity among numerous studies with divergent following criteria for studies: (i) reporting at least one correlation metric
answers on how landscape changes affect surface water quality, since that could be transformed into Pearson Correlation Coefficient (r); (ii)
complex environmental factors moderate the results. Firstly, previous only in-situ research, excluding data calculated by model simulation; (iii)
studies were conducted at small scales, and worldwide analyses have research implemented in different catchments reported in one article
been restricted to specific landscape types or water body types. Brauns should be considered as independent studies; (iv) water quality data
et al. (2022) synthesized 125 studies to reveal the human impacts on should either be obtained through parallel sampling and measured
fresh water in streams and rivers. Qiu et al. (2023) compiled 66 studies following specifications, or from authorities, both should record sam­
from around the globe on the impacts of forest cover changes on water pling locations; (v) study objects should be surface water in common
quality. Yet, so far, the responses of water quality indicators to various conditions without forcing factors (e.g. mining, fire, etc.).
landscapes at the global scale, especially the fast-degrading wetlands, To normalize the correlation metrics in the dataset, we took the
are still unclear. Secondly, previous literature at regional scales has not widely used Pearson Correlation Coefficient (r) as standard, commonly
reached a global consensus due to the strong heterogeneity in natural expressed as (Rupinski and Dunlap, 1996):
environment and socio-economic conditions worldwide. For example, ∑n
i=1 (Xi − X)(Yi − Y)
some studies concluded that agricultural development was largely r = √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
̅ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ (1)
∑n 2 ∑n
responsible for increasing the total phosphorus (TP) in water (Vaighan i=1 (Xi − X) i=1 (Yi − Y)
et al., 2017; Wei et al., 2020), while some other studies claimed TP
concentration had the strongest correlation with urban sprawl (Ren Where Xi and Yi are values of the i th independent variable and
et al., 2003; Chiang et al., 2021). Additionally, some of the study results dependent variable, and X and Y are sample means. r takes the value
could even be counterintuitive. For instance, forests might contribute to from -1 to 1. The larger the absolute value, the stronger the correlation.
the total suspended solid (TSS) output to water (Quinn & Stroud, 2002; Spearman’s Rho (ρ) and Kendall’s tau (τ) correlation coefficients
Tromboni et al., 2021), and wetland coverage could be positively related were converted to Pearson’s r through the formula (2) and (3) (Rupinski
to total nitrogen (TN) and TP concentrations (Giri et al., 2018). There­ and Dunlap, 1996).
fore, there is a necessity for a review that provides robust quantitative ( π)
syntheses across prodigious studies worldwide to comprehend the ef­ r = 2 sin ρ × (2)
6
fects of landscape changes on water quality and the variation pattern in

2
X. Shi et al. Water Research 260 (2024) 121946

Fig. 1. (a) Geographic distribution of the 625 in-situ studies worldwide included in our meta-analysis. The darker the red in the background map, the more frequently
the country is studied. All the study sites are grouped into five types according to the type of water body investigated, namely, coastal water, including coasts and
bays; flowing water, including rivers and streams; static water, including lakes, reservoirs and small ponds; watershed by geographical boundaries; and wetlands. The
amounts of studies on each of them are shown in the lower left corner. (b) Annual trends in the published papers studying the effects of landscape pattern changes on
water quality. In addition to the overall trend in dark red, the trends of the USA and China, which account for the larger share of publications, are also shown
separately from other countries. (c) Accounting for the study frequency of each water quality indicator.

r = sin(0.5 π τ) (3) mented through the Meta Package in R studio (Schwarzer, 2007). Pri­
marily, the r was converted to Fisher’s z-transformed coefficients (Zr),
The coefficient of determination (R2) of univariate linear regression which could gain better normality in distribution and calculate reliable
was equivalent to the square of r and the sign of r was the same as that of confidence intervals (Borenstein, 2009; Asuero et al., 2006), through the
the regression coefficient (Zou et al., 2003). In studies conducting following formula:
redundancy analysis (RDA) with type II scaling, the cosine of the angle ( )
between the response variable and the explanatory variable was calcu­ 1+r
Zr = 0.5ln (4)
lated to represent r (Legendre and Legendre, 2012). Data exhibited in 1− r
the figures were extracted using GetData Graph Digitizer 2.25 (GetData And accordingly, sampling variance was also derived from (Fox
Graph Digitizer, 2013). et al., 2015):
To reveal the potential influencing factors for the worldwide varia­
tion in the effects of landscape changes on water quality, environmental 1
σ 2Zr =
̂ (5)
information of those extracted studies was compiled, including the co­ n− 3
ordinates, types of water bodies, climate zones, seasons, spatial scales,
Where n is the sample size for r. Given that landscapes and water quality
and year. The majority of them were directly gleaned from the primary
in global settings generally differed, Zr within individual studies was
studies, except for the climate zones, which were allocated by co­
then accumulated by inverse variance regression and random-effects
ordinates in the Köppen-Geiger climate classification (Beck et al., 2018).
model, which was more appropriate for cases with heterogeneity
among studies than the fixed effect model (Borenstein et al., 2010). Eq.
2.2. Data Analysis (6) and (7) provided a general statement of the random-effects model:
yi = β0 + εi + ξi (6)
Meta-analysis accompanied by statistical and machine learning
methods was employed in this paper. Meta-analysis parts were imple­

3
X. Shi et al. Water Research 260 (2024) 121946

n
∑ rivers and streams (Fig. 1a). One-third studied the watersheds (herein
Y= yi ⋅wi (7) defined as a geographical area including all the water bodies rather than
i=0
a specific type), then 10 % concerned the static water bodies like lakes
Where yi was the effect size of the i th study; εi represented deviation and reservoirs. As for the water quality indicators influenced by land
within study; whereas ξi represented deviation between studies, which changes (Fig. 1c), nutrients (e.g., TP, TN and ammonia) gained major
obeyed the normal distribution with 0 as the mean and τ2 as the vari­ attention, measured in 2/3 studies, followed by physiochemical pa­
ance. We performed the restricted maximum-likelihood method (REML) rameters (e.g., DO and pH). Organic pollutants revealed by COD and
to estimate τ2 and the Q-Profile method to calculate the confidence in­ BOD were studied in nearly half of the references.
terval. Furthermore, Y in Eq. (6) was the accumulated overall effect size,
while wi representing the weight in the random-effects model was 3.2. Overall effects of landscape changes on water quality
calculated from the inverse of the total variance with τ2 as part of the
denominator. I2 was the proportion of between-study variance (τ2) in In the global context, five of the seven categories of landscape
total model variance, and in essence, indicated the heterogeneity in the compositions exhibited significant impacts on water quality indicators
overall effect after pooling (Borenstein et al., 2010). I2 less than 40 % when their coverage changed, namely urban lands, agricultural lands,
indicates negligible heterogeneity within the group of studies, and over forests, wetlands, and grasslands (Fig. 2a~e, Appendix 3). The shifts in
90 % indicates considerable heterogeneity (Deeks et al., 2022). urban land and forest coverage significantly correlated with water
Above that, it was crucial to recognize the presence of publication quality (p<0.05) and showed opposite effects with each other
bias (Nakagawa et al., 2017), which may lead to conclusions that deviate (Fig. 2a&b). There were appreciably positive relationships between
from the true results and therefore affect the validation. Hence, we urban land coverage and contaminations, including TSS (r=0.69,
generated symmetrical funnel plots and adapted Egger’s test to examine p<0.01), COD (r=0.66, p<0.01), Coliform (r=0.68, p<0.01) and Metal
the risk of bias (Nakagawa et al., 2017) (Appendix 2). ions (r=0.65, p<0.01) (Fig. 2a). Increased urban land use would also
To further investigate the moderators affecting the effects of land­ significantly reduce DO (r= -0.56, p<0.01). In contrast, increasing
scape changes on water quality, we separately established random-effect forest coverage had significantly negative correlations with those
models within subgroups for categorical variables (e.g., water body pollution indicators in water (p<0.05) and improved the DO level
types, climate zones, etc.) and adopted nonparametric statistical (r=0.50, p<0.01) (Fig. 2b). Nevertheless, the aggravating effects of
methods that greatly avoided affection by outliers and therefore reduced urban lands on contaminants were more profound than the purification
bias (Nahm, 2016). The Kruskal-Wallis test was used to inspect the effects of forests regarding TSS (rUR=0.69, rFO= -0.54), TN (rUR=0.58,
overall heterogeneity among subgroups under each moderator and the rFO= -0.44), TP (rUR=0.57, rFO= -0.53), Heavy metals (rUR=0.53, rFO=
Wilcoxon test suitable for non-independent samples was employed to -0.50) and Coliforms (rUR=0.68, rFO= -0.26) (p<0.01).
further examine heterogeneity between two-by-two subgroups since Expanding agricultural lands positively correlated with EC, Nitrate,
some samples in different subgroups were from the same site. For the TP, and DOC levels in water environment (p<0.01) (Fig. 2c). In com­
continuous variables (i.e., latitude and year), we combined the poly­ parison, the effects of agricultural land on water quality were not as
nomial regression and random-effects meta-regression (a weighted substantial as that of urban land (Fig. 2a&c), for instance, on levels of
regression model through REML), both for demonstrating variations in TSS (rUR=0.69, rAG=0.51), COD (rUR=0.66, rAG=0.38), TN (rUR=0.58,
effect sizes under moderators. rAG=0.44) and TP (rUR=0.57, rAG=0.49) (p<0.01).
The random forest model (RF) as a machine learning method was Wetland occupancy demonstrated a robust correlation with EC
adopted to find the relative importance of moderators in impacting the decreasing (r= -0.87, p<0.01) and organic matter increasing (p<0.01),
response of water quality to landscape changes via the randomForest indicated by COD(r=0.57) and DOC(r=0.56) (Fig. 2d). However, the
Package in R studio (Liaw and Wiener, 2002). Random forest regression overall correlation between wetland cover with nutrient concentrations
allowed the explanation of influences and relative importance of mul­ did not reach statistical significance at the global scale. The effects of
tiple independent variables on the dependent variable, as an ensemble grassland coverage shift on water quality were more ambiguous
classifier encompassing multitudinous decision trees (Breiman, 2001), worldwide (Fig. 2e), with a significant negative correlation with TN, TP
superior by reducing overfitting and the risk of synergy effect of inter­ and phosphate (p<0.05) and a positive correlation with microbial
dependent variables (Merghadi et al., 2020), herein the moderators. contamination (r=0.72, p<0.01). The overall effects of landscape
Random forests had their variable importance calculated using two configuration on water quality, such as the patch density changes, were
methods, of which the per cent Increase in Mean Squared Error (% also enquired and attached in Appendix 4.
IncMSE) was considered more robust, widely applied and thus employed
here (Grömping, 2009; Feng et al., 2022). A higher %IncMSE implied 3.3. Moderators influencing the landscape - water quality relationship
stronger importance of that moderator.
3.3.1. Influence of latitude and the type of water bodies
3. Results Polynomial regression curves showed that the effects of landscape
changes on water quality exhibited regularity along the latitude at the
3.1. Overview of landscape-water quality studies global scale (Fig. 3 a~c). The positive correlations of agricultural land
coverage with TN and TP were gradually enhanced with increasing
Based on the data compiled from 625 studies in 63 countries distance from the equator (p<0.01) (Fig. 3a&b). Conversely, the effects
worldwide from 1976 to 2022 (Fig. 1a&b), we could be informed that of forest coverage on TN and COD significantly weakened from low to
major studies concentrated from latitude 30◦ N to 60◦ N, mainly in the high latitudes (p<0.01) (Fig. 3a&c), despite the restricted sample
Yellow River Basin and the Yangtze River Basin, China, the Mississippi numbers in the southern hemisphere. Regarding the urban land
River Basin and the Great Lake Region, USA and various basins in coverage, its aggravating contribution to TP and COD levels perceptibly
Europe (Fig. 1a). Studies on the relationship between landscape changes declined as the latitude rose (p<0.05 and p<0.01, respectively)
and water quality were almost in Northern America and the Europe (Fig. 3b&c). The effects of wetland coverage on TN, TP and COD levels
before 2007. The years 2008 and 2020 witnessed rapid growth in study were not significantly associated with latitude (p>0.05) (Fig. 3a~c).
quantities, especially in China (Fig. 1b). To date, there have been large Significant variations across different water body types were wit­
gaps in research in Africa and Northern Asia. nessed in the correlations between landscape changes and water quality
Around half of the studies investigated flowing water bodies like (Fig. 3d~e). Among all the landscape compositions, the influence of
agricultural land changes on water quality exhibited the most significant

4
X. Shi et al. Water Research 260 (2024) 121946

Fig. 2. Overall effects of Individual landscape composition on each water quality indicator under the random effects model. UR= Urban lands, AG = Agricultural
lands, FO = Forests, WET= Wetlands, GR =Grasslands. Different colors for distinguishing water quality parameters: blue for physiochemical parameters, yellow for
solids, purple for nutrients, green for organic pollutants, orange for biological pollutants, and magenta for metal cations. Bold font represents statistically significant
correlations at p < 0.05. ‘n’ denotes the number of sampling data. Squares with error bars denote the overall correlation coefficients (CC) (r) and the 95 % confidence
interval (CI). The greater the CC is above 0 indicates the stronger the positive relation is, and vice versa. The size of the square symbolizes the effect size. Note: ‘*’
denotes p<0.05, ‘**’ denotes p<0.01, ‘***’ denotes p<0.001.

heterogeneity in different water body types (p<0.05) (Fig. 3d~e), of efficient purification effect on TP in coastal water (rcoasts= -0.68,
which the weakest influence was always exerted on water in watersheds p=0.02) (I2=0). Urban expansion had the most profound impact on COD
rather than a specific type (Fig. 3d~e). Remarkably, the elevating effects level in lake water (rlakes= 0.87, p<0.01). When wetlands were regarded
of agricultural land on TN (rsteams=0.64, rrivers=0.54) and TP as a type of water body, they showed unique within-group response
(rsteams=0.74, rrivers=0.63) were stronger in streams than in rivers consistency globally (mostly I2=0) under landscape changes (Fig. 3d~f).
(p<0.05) (Fig. 3d&e). As the coverage of urban lands or forests
changed, their weakest influences on water quality were mainly
observed in reservoirs (p<0.05) (Fig. 3e&f). Forests exerted the most

5
X. Shi et al. Water Research 260 (2024) 121946

Fig. 3. Correlation coefficients between landscape compositions (AG, UR, FO and WET) and water quality parameters respectively TN (a, d), TP (b, e) and COD (c, f)
moderated by latitude (a~c), and type of water bodies (d~f). (a~c) exhibit the polynomial regression curves for the correlation coefficient (r) with 95 % CI under
latitude changes, the adjusted R2 (R2adj) and significance (P) of which is also reported. (d~f) shows the correlation coefficient (r) within each water body type
subgroup and the heterogeneity between subgroups by random-effect models, in which error bars denote the 95 % CI. Percentages adjacent to error bars are the I2
value which is less than 40 % and indicates the heterogeneity within group could be ignored (Deeks et al.,2022). P-values for Kruskal- Wallis test are also reported
beneath and those significant figures are bolded. Note: ‘*’ denotes p<0.05, ‘**’ denotes p<0.01, ‘***’ denotes p<0.001.

3.3.2. Difference between seasons and climate zones in effects of landscape (Table 1, Fig. 4). The strongest effect size of forest on TN was rdry= -0.73,
on water quality on TP was rdry=-0.70 and on COD was rdry= -0.67 (p<0.01), all
Climatic conditions could influence the temperature, precipitation, appearing in the dry season. Similarly, the dry season witnessed the
etc., thus making the relationship between landscape and water quality strongest boosting of agricultural land on TN (rdry=0.65), TP (rdry=0.72)
different among seasons and climate zones. Seasonally, the effects of and COD (rdry=0.82) (p<0.01). The influence of urban expansion on
landscape changes on TN and TP concentration were more powerful in COD was rising as the season became moister, from rdry=0.50 to rnormal=
the dry season and the rainy season than in normal period (p<0.01) 0.62 to rwet= 0.68 (p<0.001). While the effects of wetland coverage
change on water were not dependent on seasons (Table 1, Fig. 4).
Table 1 Climate zone could be a pronounced moderator under the urban land
Kruskal-Wallis test p-value indicating the significance of the difference in the changes (p<0.01) (Table 1). Expanding urban land coverage had the
correlation coefficients contributed by the various moderators. strongest impact on TN (rarid=0.71), TP (rarid=0.94) concentration and
COD (rarid=0.70) in the arid area among all the climate zones (p<0.05)
Moderators TN TP COD
(Fig. 4). Although climatic zone was not as significant a moderator in the
Seasonality
relationship between agricultural land changes and water quality as
WET 0.17 0.28 0.97
UR <0.01 ** <0.01 ** <0.01 ** they were for urban land, the similar severity in arid areas also occurred
FO <0.01 ** <0.01 ** 0.76 with rarid=0.78 for TN, rarid=0.86 for TP and rarid=0.83 for COD
AG <0.01 ** <0.01 ** 0.16 (p<0.01). Besides, the elevating effects of wetlands on COD were
Climate zone significantly influenced by climate zones (p<0.01) (Table 1), peaking in
WET 0.64 0.28 <0.01 **
UR <0.01 ** <0.01 ** <0.01 **
cold areas (rcold=0.82, p<0.001).
FO 0.73 0.12 0.23
AG 0.24 0.06 0.10 3.3.3. Spatial and temporal changes in the landscape - water quality
Spatial scale relationship
WET 0.38 0.59 0.80
The effects of landscape changes on water quality were scale-
UR 0.11 0.01 * 0.12
FO 0.22 0.03 * 1.00 dependent both at spatial and temporal scales. When the geographic
AG <0.01 ** <0.01 ** 0.10 scale of investigation was adjusted, the effects of agricultural land on
nutrient concentrations experienced the most apparent variation
Note: ‘*’ denotes p<0.05, ‘**’ denotes p<0.01.

6
X. Shi et al. Water Research 260 (2024) 121946

Fig. 4. Summary of correlation coefficients between landscape compositions (AG, UR, FO and WET) and water quality parameters respectively TN, TP and COD
moderated by seasonality, climate zone and scale under the random effects model. ‘*’ denotes p<0.05, ‘**’ denotes p<0.01, ‘***’ denotes p<0.001. The square with
error bar denotes the correlation coefficients (r) and 95 % CI for each subgroup. The greater the CC is above 0 indicates the stronger the positive relation is, and vice
versa. The size of the square symbolizes the effect size. The I2 value represents the heterogeneity within individual subgroups in the random effects model. Smaller I2
implies less heterogeneity within the subgroup. The W-p significance is from the Wilcoxon test for the disparity between each subgroup and a certain subgroup of
their moderator (i.e., Annual of the Seasonality, Temperate of the Climate zone, and Catchment of the Scale).

(p<0.01) (Table 1). There was a tendency in Fig. 4 that as the buffer (Table 1).
radius expanded outwards along the water body boundaries, the cor­ From the perspective of temporal evolution, the meta-regression
relation coefficients between agricultural land coverage and water lines exhibited an intensifying tendency in the impacts of all four
quality parameters decreased until 500-1000m and then climbed to the landscape composition changes on water quality from pre-1990 to 2022
maximum when the buffer was over 2000m (W-p<0.05). Similarly, (Fig. 5a~c). The correlations of agricultural land changes with TN and
buffer scales significantly modulated the effects of forest and urban land TP showed a concave trend through time, leading to a smaller effect size
on TP concentration (p<0.05) (Table 1), in which a similar decreasing in the 2020s than in the 1990s (p<0.01) (Fig. 5a&b). Studies on water
and then increasing tendency of effect size as the scale broadened also quality response to land changes in urban areas and forests emerged
held. Further, the impacts of the above landscape changes on water around 1995, later than those in agricultural uses. The temporal varia­
environment at larger scales tended to witness a global consistency, i.e., tion of urban land and forest effects on water quality appeared syn­
at the catchment scale (p<0.05) or the >2000m buffer scale (p<0.001) chronized, whether along the linear growth (for TN and TP) or
(Fig. 4). In addition, responses of water quality to wetland coverage fluctuating growth (for COD). In particular, all the growth rates of the
changes did not show significant differences across diverse spatial scales effects of forest were higher than those of urban land (p<0.05). Their

7
X. Shi et al. Water Research 260 (2024) 121946

coupled trends might result from their dominantly opposite effects

(Fig. 2a&b), i.e., forests could perform the purification capacity to a
greater extent as the pollution intensity of urban land rose when the
pollutant concentrations were not saturated. As for wetland coverage
changes worldwide, its effects on water quality were not significantly
changed over time (p>0.05) (Fig. 5a~c).

3.4. Importance analyses of moderators to water quality variables

The random forest regression furnished the orders of the relative

importance of moderating variables above (Fig. 6). Latitude most
significantly influenced the correlation between agricultural land and
water quality, in which scale and year were also among the critical
moderators (Fig. 6a). It could be generally concluded that climatic
conditions exerted less significant adjustments on how agricultural
coverage changed water quality (Fig. 6a~c). Whereas seasonality had a
prominent role in the response of water quality indicators to changes in
riparian urban lands (Fig. 6d~f). Moreover, the rankings demonstrated
that latitude was way ahead as an important explanatory factor for the
variation of forest impacts on TN, TP and COD (Fig. 6g~i). Compared
with the three landscape components above, the effects of wetland on
water quality tended to be dependent on climate zone (Fig. 6h&i).

4. Discussion

4.1. Study trends and hotspots

Our global review revealed that the shifting of research centers of the
landscape-induced water quality changes, as shown in Fig. 1a, was fully
correlated with the regional socioeconomic development and urbani­
zation process through time (Sun et al., 2020; Radwan et al., 2021). The
quantitative research on the correlation between agricultural occupa­
tion and water quality started earlier than the other landscape types as
shown in Fig. 5, echoing a series of policies such as the Clean Water Act
in the USA in 1972 to control NonPoint Source pollution (NPS) (Haith,
1976; Mansaray et al., 2018). As demonstrated in Fig. 1, chemical,
bacterial, and sediment loadings from NPS, mostly resulting from land
runoff and hard to detect, had subsequently been a sustained academic
concern and the USA dominated the studies (Wilkin and Jackson, 1984;
Dauer et al., 2000; Mehaffey et al., 2005). The Pollution Prevention Act
launched in the USA in 1990 and the Law of the People’s Republic of
China on the Prevention and Control of Water Pollution in 1996 might
be part of the drivers of emerging studies on the effects of urban lands
and forests on water quality around 1996 as shown in Fig. 5, matching
the upward trend of study amounts. With the increasing worries about
water quality changes under human activities and climate changes after
2000, the research on their dynamic relationship was universally
enriched (Fig. 1), integrating methods such as the Better Assessment of
Integrating Point and Non-point Sources (BASINS) framework (Bhat­
tarai et al., 2008), ArcView Generalized Watershed Loading Function
(AVGWLF) model (Tu, 2009), and other conceptual models for simu­
lating future changes under different scenarios (Erol and Randhir,
2013). The emphasis in the Chinese Government’s 11th Five-Year Plan
(2006-2010) on mitigating climate change and strengthening pro­
jections for pollutant additions could have stimulated the exponential
Fig. 5. Temporal changes of the correlation coefficients between landscape growth of China’s research in 2008 (Xu et al., 2019), as illustrated in
compositions (AG, UR, FO and WET) and water quality parameters respectively Fig. 1b. Attention to relevant issues globally spiked again around 2020,
(a) TN, (b) TP and (c) COD. The solid line with 95 %CI denotes the relation of which studies about the impacts of LULC changes on water quality
curve fitted by polynomial regression, and the adjusted R2 (R2adj) and signifi­ reached new levels in developing countries in Africa and South America,
cance (P) of which are also reported. The dotted line denotes the relation curve for instance, Bolivia (Gossweiler et al., 2019), Cameroon (Ewane, 2020),
fitted by random-effects meta-regression with the circle representing sample
Ethiopia (Woldeab et al, 2019), Ghana (Gyimah et al, 2020), Uruguay
data, whose size is proportional to the weight of effect size. Note: ‘*’ denotes
(Gorgoglione et al, 2020), and Zambia (Winton et al, 2021). Nonethe­
p<0.05, ‘**’ denotes p<0.01, ‘***’ denotes p<0.001.
less, as revealed in our map (Fig. 1a), some places were still vacant
without studies, even if facing the intensifying water quality challenges
caused by landscape changes (Fig. 5).

8
X. Shi et al. Water Research 260 (2024) 121946

Fig. 6. Relative importance rank of various factors on variation in effects of changes in the landscape (AG, UR, FO and WET) on water quality, represented by the
impacts on TN (a,d,g,j), TP (b,e,h,k), and COD (c,f,i,l). Importance decreases from top to bottom. The R2 of random forest model is also reported separately in the
figure. Fig. 6 Relative importance rank of various factors on variation in effects of changes in the landscape (AG, UR, FO and WET) on water quality, represented by
the impacts on TN (a,d,g,j), TP (b,e,h,k), and COD (c,f,i,l). Importance decreases from top to bottom. The R2 of random forest model is also reported separately in
the figure.

4.2. Effects of agricultural lands on water quality radiuses (Fig. 4).

Apart from the changes in effect size, we found that agricultural land
Our meta-analysis results substantiated that increasing agricultural could show opposite effects on water quality at different scales at the
land coverage in the catchment actively contributed to nutrient pollu­ first time (Fig. 4), which universally presented a negative correlation
tion and solids concentration in water worldwide (Fig. 2c). A consid­ within 500-1000m proximity riparian. A possible explanation was that
erable amount of nutrients that were not fully utilized in fertilizer and other landscape compositions might cause compelling impacts that
feed inputs in modern intensive farming and livestock farming entered overpower the positive effect of agricultural land on water pollution. For
the water with irrigation water and surface runoff, enriching N and P in instance, when urbanized spaces increased faster or in highly urbanized
the water and causing harmful eutrophication (Woli et al., 2004; Khan areas as in studies by Tu and Xia (2008) and Zhao et al. (2015), the
and Mohammad, 2014). Additionally, agricultural activities such as existence of agricultural land might be obscured and presented a nega­
ploughing and irrigation undeniably caused water erosion and increased tive correlation with water quality deterioration. Another factor we
both suspended and dissolved solids in water (Chen et al., 2017). At the could take into interpretation was the endogenous characteristics of the
same time, the nutrients in particulate matter, especially phosphorus, agricultural lands. It has been reported in several studies that paddy,
would be further released into the aquatic environment (Horppila, instead of other cropland, might have a capacity for pollution reduction
2019). by strong plant absorption stimulated by appropriate environmental
The spatial-scale effect was witnessed in the nonpoint-source pollu­ conditions in growth (Jung et al., 2008; Zhang et al., 2010). However,
tion exported from agricultural land (Table 1, Fig. 6) since distance due to the substantial heterogeneity in geographical environment be­
greatly affected the transport and transformation processes of the tween watersheds, any future study on a single scale should be cautious
abovementioned pollutants (Sliva and Williams, 2001; Tu and Xia, and the threshold effects in the impacts of agricultural land coverage on
2008). Our global meta-analysis settled the long-term dispute and water quality are expected for further investigation.
informed that the most powerful influencing scales of agricultural land
for water quality were the whole catchment and over 2000m buffer scale
since they had no statistical differences (Fig. 4). Before our study, the 4.3. Response of water quality to urbanization
200m buffer scale (Tran et al., 2010), 500m buffer scale (Gove et al.,
2001) and catchment scale (Sliva and Williams, 2001; Zhang et al., Our synthesized analysis proclaimed that contaminant increase and
2019) were previously identified as the most significant scales in DO decrease in surface water caused by urban expansion have persisted
influencing water quality when landscape changed within watersheds. worldwide for decades (Fig. 2a). Continuous impervious lands caused
Further, our results indicated that study results worldwide held the massive nonpoint-source pollution by surface runoff exporting solids,
largest consistency at the catchment scale compared with other buffer nutrients, and particularly, organic, bacterial, and heavy metal pollut­
ants (Miller and Hutchins, 2017; Viau et al., 2011; Kang et al., 2010).

9
X. Shi et al. Water Research 260 (2024) 121946

Concentrated pollution from urban lands could alter the processes of 4.4. Correlation of forest with water quality
regional biogeochemical cycling and consume DO in water (Tong et al.,
2020). Further, compared with agricultural landscapes, the urban Our meta-analysis concluded that forest cover significantly
environment could increase regional rainfall and thus boost the washing decreased the risk of water pollution (Fig. 2b), especially the biological
and leaching of contaminants from dust, food remnants and maintained and heavy metal contamination, suggesting the importance of forest
greenspaces (Liu and Niyogi, 2019; Morée et al., 2013; Simpson et al., restoration in global urbanization for water security. Forests could cool
2022). Considering that, our meta-analysis indicated that nutrients like water flow and therefore increase the solubility of oxygen in water
TN and TP related more strongly to urban lands than agricultural pro­ (Garner et al., 2014), which was necessary for the aquatic organism
ductions (Fig. 2aandc). respiration and the biological and biochemical decomposition of organic
Our study further demonstrated that the worldwide correlation matter (Ansa-Asare, 2000). Moreover, given that (i) tree canopy was key
variation of water quality with urban land coverage was dramatically to stormwater management and soil erosion regulation, (ii) the inter­
related to climate conditions, especially seasons (Table 1, Fig. 6e~f). ception by tree roots and litters, (iii) the absorption of bioavailable
The overall impacts of urban lands on water quality in both wet and dry contaminants by robust roots, the forest could cogently prevent the
seasons were more severe than those in normal seasons with distinct solids and nutrients from transporting into water (Lowrance et al., 1997;
mechanisms (Fig. 4). The wet season represented greater surface runoff Quinn and Stroud, 2002; Harris, 2001).
and thus more pollutants washed into water (Regier et al., 2020), the Latitude ranked top as the contributing factor in changes in the
peak of which in extreme precipitation could be even worse than point correlation of forest with water quality (Fig. 6). The low-latitude forests
source pollution according to Pak et al. (2021). Thus, the pollutant were the most efficient at water purification (Figs. 3 and 6). The benefits
discharge from urban areas could exceed the self-purification capacity of of forest cover for water quality seemed to gradually decrease as the
water bodies and deteriorate water quality (Lai et al., 2013). In the dry latitude increases (Fig. 3). The biomass of forest was considered as an
season, declining water volumes would diminish the dilution towards underlying contributor in this pattern since observations demonstrated
contaminants and the self-purification capacity of water environment, that the highest aboveground biomass of forests occurred in tropical
leading to water quality decline (Xiao et al., 2016; Bussi et al., 2017). forests and then the northern temperate forests (Schepaschenko et al.,
Regarding TP, our result showed the dry season could amplify its cor­ 2019). Another potential explanation was that the high temperature and
relation with urban land than the wet season (Fig. 4), which might result evapotranspiration increased the solubility of bioavailable contaminants
from the concentration of point-source pollutants and the release of in water (such as phosphorus) ((Yang et al., 2023; Cheng et al., 2020)
particulate phosphorus from sediment when water levels decreasing and accelerated chemical and biochemical processes (such as denitrifi­
(Bussi et al., 2017; Record et al., 2016). In addition, our study found that cation and carbon released from organic matter decomposition) (Daw­
globally, dry and wet seasons amplified TN output from urban area to a son and Murphy, 1972; Aerts, 1997). However, the tropical forests have
similar extent (Fig. 4), answering the divergent in regional studies. Some been suffering from deforestation, particularly in those underdeveloped
believed that nitrogen tended to be deposited in surface soils in the dry countries relying the trade (Hoang and Kanemoto, 2021; Zhang and Wei,
season with poor hydrological connectivity, reducing N export from 2021), which called on global sharing sustainable development and
urban lands into receiving water (Ferrier et al., 1995; Wu et al., 2022). water security responsibility.
However, other scholars demonstrated that the dry season could amplify Additionally, hinted by outliers in our global analysis, some excep­
the pollution from urban lands upon TN (Ding et al., 2015; Zhang et al., tions need attention in the water purification functions of forests. Firstly,
2019). Additionally, our result witnessed an increasing effect of urban hilly areas and flooded areas seemed to promote the probability of forest
lands on COD as precipitation rose (Fig. 4), considered to be related to TSS, TN and TP output owing to the erosion of topsoils (Quinn and
increasing runoff, especially in slightly urbanized catchments without Stroud, 2002; Tram et al., 2022; Tromboni, 2021). Secondly, the nega­
complete drainage systems (Liu et al., 2017; Nafi’Shehab et al., 2021), tive connection between forest coverage and DO observed in the outliers
and as comparison highly-urbanize catchment might observe higher might be elucidated by two reasons: excess organic matter input from
COD in dry seasons (Chen et al., 2016; Liu et al., 2017). litter, especially in low-flow periods (Abdul-Aziz and Ahmed, 2017); and
Furthermore, as demonstrated in Fig. 4, arid areas tend to suffer the static wind and sequent still water status caused by enclosed forest,
more harsh deterioration in water quality as urban land sprawl world­ which was not conducive for oxygen exchange (Scully, 2010; Chen et al.,
wide, taking New Mexico in the USA (Regier et al., 2020) and Xinjiang in 2021a). Additionally, unfragmented forest patches with complex edge
China (Wang et al., 2022b) as examples. Low quality and quantity of shapes could increase the benefits of water purification, as supported by
water availability would exacerbate water scarcity in arid regions, studies in Brazil (de Mello et al., 2020), Malaysia (Nafi’Shehab et al.,
exacerbating inequalities in global water scarcity and requiring urgent 2021), China (Yu et al., 2013), Korea (Lee et al., 2009), and the United
attention (Wang et al., 2024). States (Carey et al., 2011).
Combined with spreading hotspots on the impacts of urban sprawl on
the water environment in recent years (Fig. 1), we found that rapid 4.5. Connection of wetlands with water quality
urbanization in developing regions in arid and tropic areas, especially
those without sound drainage and sewage treatment systems and prone Although numerous ecological experiments in small scales proved
to stormwater flooding hazards from concentrated precipitation, could the role of wetlands in removing pollutants from water, our global meta-
lead to severe pathogens contamination (mainly fecal coliform) and take analysis showed that changes in wetland coverage in watershed did not
thousands of lives through infectious disease spread (Ashbolt,2004). have a significant impact on water quality, which might be because the
Examples include the pollution in the Mun River watershed in Thailand coverage of wetland was rather small in the whole landscape (Fig. 2d).
(Yadav,2019), the Mekong tributary watersheds in Lao P.D.R (Ribolzi For mechanism, wetlands achieved water purification mainly through
et al.,2011) and the Intag area in Ecualor (Knee and Encalada, 2014). As physicochemical processes represented by sedimentation and adsorp­
a comparison, the water quality had improved through industrial tion and biochemical processes represented by wetland plants’ uptake
restructuring and water management in some developed regions, taking and soil microorganisms’ degradation (Verhoeven et al., 2006). In
Lake Ontario in Canada (Croft-White et al., 2017) and North Canal River anaerobic wetland environments, denitrifying bacteria sequentially
in China as examples (Zhu et al., 2023). The imbalance in converted inorganic nitrogen into nitrogen gas, thus removing nitrogen
socio-economic development aggravated the inequity in global water from the water (Gersberg, 1986; Verhoeven et al., 2006; Morrissy et al.,
security. 2021). Apart from being directly absorbed by wetland plants (Vymazal,
2007), phosphorus was easily bound and deposited in the wetland soil
through a series of complex reactions with metal ions in the sediments

10
X. Shi et al. Water Research 260 (2024) 121946

(Records et al., 2016). The decreasing effects on TN, nitrate and TP could and temporal resolution at more ambitious research scales at low cost
be partly observed in the overall effects of wetlands in Fig. 2d but were (Ramadas and Samantaray, 2018). As monitoring and prediction models
very ambiguous (cross the 0 line). The mist came from some references of climate changes and hydroclimatic extremes continue to be refined,
reflecting the wetland coverage positively correlated with nutrient future research is expected to integrate the drivers of water quality
concentration (Rothenberger, 2009; Giri et al., 2018; Bu et al., 2014). changes considering changes in climate and landscape patterns and
We surmised that one reason could be the nitrogen fixation by wetland other factors such as hydrological processes and human activities (van
plants converting airborne nitrogen to reactive nitrogen in water (Jor­ Vliet et al., 2023).
dan, 2011). Additionally, an explanatory phenomenon was that the
wetland could absorb or release nutrients in water depending on how 5. Conclusions
much the concentration was and whether it was saturated or not (Day
et al., 2004; Verhoeven et al., 2006). Previous studies indicated the Taking advantage of meta-analysis and a fusion of various statistical
existence of critical loads, whereby both community structure and and machine learning methods, our study compiled the global effects of
ecological functions, including water purification, would be shifted landscape changes on water quality. Results revealed the expansion of
upon reaching a certain level of nutrient concentrations in wetland urban land was most responsible for the deterioration of water quality,
aquatic environments (Verhoeven et al., 2006). Hence, intense nitrogen more so than agricultural land even in nutrient pollution. Forest
and phosphorus inputs from surrounding urban and agricultural lands coverage in watersheds generally exerts improvement in water quality.
exceeding the loads would lead the wetland to behave as a nutrient The intuitive water purification function of wetlands is obscured in
exporter. mixed landscape studies, but wetlands are significantly and positively
Our meta-analysis revealed COD and DOC in water significantly correlated with organic matter-related indicators. The effects of those
positively responded to wetland coverage worldwide (Fig. 2d), typically landscape changes on water quality are moderated to varying degrees by
observed in studies in Canada (Chen et al., 2021b), the USA (Tu, 2011), factors such as latitude, water body type, seasonality, climatic zone, and
and China (Bu et al., 2014). Owing to the flooded and oxygen-deficient spatial scale, and gradually increase over time. The impacts of surface
state of the wetland, the proportion of organic carbon decomposition land change on water environment quality are such a universal and
was always limited (Chen, 2023). Coupled with captured incoming growing problem that sustainable LULC management and enhanced
biomass and organic matter, wetlands could sequester and store carbon water protection through nature-based solutions are urgently needed.
in soil (Yu et al., 2021), performing as a carbon sink (Freeman et al., This paper also revealed the regional unevenness in the studies and
2004). Increasing temperature and future warming would enhance the appealed to further attention to global water equity and environmental
correlation between wetlands and carbon concentration in water by justice under intensified climate changes.
accelerating the organic carbon leaching from soil texture and
increasing the organic matter in water (Freeman et al., 2004; Evans CRediT authorship contribution statement
et al., 2005).
The correlation of wetland coverage with COD levels in water Xinying Shi: Writing – original draft, Visualization, Methodology,
significantly differed among climatic zones and latitudes (Table 1, Investigation, Formal analysis, Data curation. Dehua Mao: Writing –
Fig. 6). Cold areas and high-latitude regions witnessed a distinctively review & editing, Visualization, Validation, Funding acquisition,
high effect size (Fig. 4), which was highly coherent with the previous Conceptualization. Kaishan Song: Writing – review & editing, Valida­
findings in boreal peatlands by Pastor et al. (2003) and the global tion. Hengxing Xiang: Writing – review & editing. Sijia Li: Writing –
variation patterns of nitrogen revealed by Jordan (2011). That could review & editing. Zongming Wang: Writing – review & editing,
result from the long freezing period and long water residence in the cold Supervision.
high-latitude environment. Besides, low temperatures generally
restricted microbial activity and chemical reaction rates, reducing Declaration of competing interest
organic matter decomposition efficiency (Andersson, 2000), DOC
catabolism efficiency, and nitrification and denitrification efficiency etc. The authors declare that they have no known competing financial
(Evans et al., 2005; Morrissy et al., 2021), and thus increased the total interests or personal relationships that could have appeared to influence
concentration of organic carbon in water. Other moderators (season­ the work reported in this paper.
ality, spatial scale, year, etc.) hardly showed influence on the effects of
wetland coverage changes on water quality in our analysis (Table 1, Data availability
Figs. 4-6), perhaps related to their self-regulating capacity.
Data will be made available on request.
4.5. Limitations and future directions

Although this study revealed the global effects of landscape changes Acknowledgements
on surface water quality based on extensive studies, sample size and
distribution still majorly restricted this data-driven meta-analysis. This work was supported by the National Natural Science Foundation
Wetlands did not receive that much attention in the compositive land­ of China (U23A2008 and 42171379), the funding from Youth Innova­
scape studies compared with other landscape compositions (Fig. 2a~e). tion Promotion Association of CAS (2017277), and the Young Scientist
The previous studies in the developing countries around the equator and Group Project of Northeast Institute of Geography and Agroecology,
the southern hemisphere were relatively sparse (Fig. 3), calling for more Chinese Academy of Sciences (2022QNXZ03).
stakeholders caring for the existing and potential water quality issues
under the landscape changes in those distribution gaps. Supplementary materials
Future research should follow technological and methodological
developments. High-frequency in-situ water quality monitoring tech­ Supplementary material associated with this article can be found, in
nology such as the Online-monitoring Systems (Ministry of Ecology and the online version, at doi:10.1016/j.watres.2024.121946.
Environment, 2019) can effectively improve the temporal continuity in
water sampling and the standardization of water quality measurement.
Hyperspectral remote sensing technology combined with machine
learning algorithms contributes to water quality data with high spatial

11
X. Shi et al. Water Research 260 (2024) 121946

References Delpla, I., Rodriguez, M.J., 2014. Effects of future climate and land use scenarios on
riverine source water quality. Sci. Total Environ. 493, 1014–1024.
Ding, J., Jiang, Y., Fu, L., Liu, Q., Peng, Q., Kang, M., 2015. Impacts of land use on
Abdul-Aziz, O.I., Ahmed, S., 2017. Relative linkages of stream water quality and
surface water quality in a subtropical River Basin: a case study of the Dongjiang
environmental health with the land use and hydrologic drivers in the coastal-urban
River Basin, Southeastern China. Water. (Basel) 7 (8), 4427–4445.
watersheds of southeast Florida. GeoHealth 1 (4), 180–195.
Erol, A., Randhir, T.O., 2013. Watershed ecosystem modeling of land-use impacts on
Aerts, R., 1997. Climate, leaf litter chemistry and leaf litter decomposition in terrestrial
water quality. Ecol. Modell. 270, 54–63.
ecosystems: a triangular relationship. Oikos. 439–449.
Evans, C.D., Monteith, D.T., Cooper, D.M., 2005. Long-term increases in surface water
Andersson, S., Nilsson, S.I., Saetre, P., 2000. Leaching of dissolved organic carbon (DOC)
dissolved organic carbon: observations, possible causes and environmental impacts.
and dissolved organic nitrogen (DON) in mor humus as affected by temperature and
Environ. Pollut. 137 (1), 55–71.
pH. Soil Biol. Biochem. 32 (1), 1–10.
Ewane, E.B., 2020. Assessing land use and landscape factors as determinants of water
Ansa-Asare, O.D., Marr, I.L., Cresser, M.S., 2000. Evaluation of modelled and measured
quality trends in Nyong River basin, Cameroon. Environ. Monit. Assess. 192 (8), 507.
patterns of dissolved oxygen in a freshwater lake as an indicator of the presence of
Feng, Z., Wang, L., Wan, X., Yang, J., Peng, Q., Liang, T., Rinklebe, J., 2022. Responses of
biodegradable organic pollution. Water. Res. 34 (4), 1079–1088.
soil greenhouse gas emissions to land use conversion and reversion—a global meta-
Ashbolt, N.J., 2004. Microbial contamination of drinking water and disease outcomes in
analysis. Glob. Chang. Biol. 28 (22), 6665–6678.
developing regions. Toxicology. 198 (1-3), 229–238.
Ferrier, R.C., Whitehead, P.G., Sefton, C., Edwards, A.C., Pugh, K., 1995. Modelling
Asuero, A.G., Sayago, A., González, A.G., 2006. The Correlation Coefficient: An
impacts of land use change and climate change on nitrate-nitrogen in the River Don,
Overview. Crit. Rev. Anal. Chem. 36 (1), 41–59. https://doi.org/10.1080/
North East Scotland. Water. Res. 29 (8), 1950–1956.
10408340500526766.
Flood, M.T., Hernandez-Suarez, J.S., Nejadhashemi, A.P., Martin, S.L., Hyndman, D.,
Beck, H.E., et al., 2018. Present and future Köppen-Geiger climate classification maps at
Rose, J.B., 2022. Connecting microbial, nutrient, physiochemical, and land use
1-km resolution. Sci. Data 5 (1), 1–12.
variables for the evaluation of water quality within mixed use watersheds. Water.
Bhattarai, G., Srivastava, P., Marzen, L., Hite, D., Hatch, U., 2008. Assessment of
Res. 219, 118526.
economic and water quality impacts of land use change using a simple bioeconomic
Fluet-Chouinard, E., Stocker, B.D., Zhang, Z., Malhotra, A., Melton, J.R., Poulter, B.,
model. Environ. Manage 42, 122–131.
McIntyre, P.B., 2023. Extensive global wetland loss over the past three centuries.
Borenstein, M., Hedges, L.V., Higgins, J.P., Rothstein, H.R., 2009. Introduction to meta-
Nature 614 (7947), 281–286.
analysis. Wiley, Chichester.
Fox, G.A., Negrete-Yankelevich, S., Sosa, V.J. (Eds.), 2015. Ecological Statistics:
Borenstein, M., Hedges, L.V., Higgins, J.P., Rothstein, H.R., 2010. A basic introduction to
Contemporary Theory and Application. Oxford University Press.
fixed-effect and random-effects models for meta-analysis. Res. Synth. Methods 1 (2),
Freeman, C., Fenner, N., Ostle, N.J., Kang, H., Dowrick, D.J., Reynolds, B., Hudson, J.,
97–111.
2004. Export of dissolved organic carbon from peatlands under elevated carbon
Brauns, M., Allen, D.C., Boëchat, I.G., Cross, W.F., Ferreira, V., Graeber, D., Gücker, B.,
dioxide levels. Nature 430 (6996), 195–198.
2022. A global synthesis of human impacts on the multifunctionality of streams and
Galbraith, L.M., Burns, C.W., 2007. Linking land-use, water body type and water quality
rivers. Glob. Chang. Biol. 28 (16), 4783–4793.
in southern New Zealand. Landsc. Ecol. 22, 231–241.
Breiman, L., 2001. Random forests. Mach. Learn. 45, 5–32.
Garner, G., Malcolm, I.A., Sadler, J.P., Hannah, D.M., 2014. What causes cooling water
Broussard, W., Turner, R.E., 2009. A century of changing land-use and water-quality
temperature gradients in a forested stream reach? Hydrol. Earth. Syst. Sci. 18 (12),
relationships in the continental US. Front. Ecol. Environ. 7 (6), 302–307.
5361–5376.
Bu, H., Meng, W., Zhang, Y., Wan, J., 2014. Relationships between land use patterns and
Gersberg, R.M., Elkins, B.V., Lyon, S.R., Goldman, C.R., 1986. Role of aquatic plants in
water quality in the Taizi River basin, China. Ecol. Indic. 41, 187–197.
wastewater treatment by artificial wetlands. Water. Res. 20 (3), 363–368.
Bullard, W.E., 1966. Effects of land use on water resources. J. (Water Pollut. Control
GetData Graph Digitizer, 2013. GetData graph digitizer (2.25). Computer Software.
Federat.) 38 (4), 645–659. http://www.jstor.org/stable/25035540.
Giri, S., Qiu, Z., Zhang, Z., 2018. Assessing the impacts of land use on downstream water
Bussi, G., Janes, V., Whitehead, P.G., Dadson, S.J., Holman, I.P., 2017. Dynamic response
quality using a hydrologically sensitive area concept. J. Environ. Manage 213,
of land use and river nutrient concentration to long-term climatic changes. Sci. Total
309–319.
Environ. 590, 818–831.
Gorgoglione, A., Gregorio, J., Rios, A., Alonso, J., Chreties, C., Fossati, M., 2020.
Carey, R.O., Migliaccio, K.W., Li, Y., Schaffer, B., Kiker, G.A., Brown, M.T., 2011. Land
Influence of land use/land cover on surface-water quality of Santa Lucìa river,
use disturbance indicators and water quality variability in the Biscayne Bay
Uruguay. Sustainability. 12 (11), 4692.
Watershed, Florida. Ecol. Indic. 11 (5), 1093–1104.
Gossweiler, B., Wesström, I., Messing, I., Romero, A.M., Joel, A., 2019. Spatial and
Carpenter, S.R., Caraco, N.F., Correll, D.L., Howarth, R.W., Sharpley, A.N., Smith, V.H.,
temporal variations in water quality and land use in a semi-arid catchment in
1998. Nonpoint pollution of surface waters with phosphorus and nitrogen. Ecol.
Bolivia. Water. (Basel) 11 (11), 2227.
Appl. 8 (3), 559–568.
Gove, N.E., Edwards, R.T., Conquest, L.L., 2001. Effects of scale on land use and water
Chen, D., Wei, W., Chen, L., 2017. Effects of terracing practices on water erosion control
quality relationships: a longitudinal basin-wide perspective. JAWRA J. Am. Water
in China: A meta-analysis. Earth. Sci. Rev. 173, 109–121.
Res. Assoc. 37 (6), 1721–1734.
Chen, H., Yu, Z., Wu, N., Wang, Y., Liu, X., 2023. C-wetland”: a new paradigm to enhance
Grömping, U., 2009. Variable importance assessment in regression: linear regression
conservation of carbon-rich wetlands. The Innovation 4 (2), 100403.
versus random forest. Am. Stat. 63 (4), 308–319.
Chen, Q., Mei, K., Dahlgren, R.A., Wang, T., Gong, J., Zhang, M., 2016. Impacts of land
Gyimah, R.A.A., Karikari, A.Y., Gyamfi, C., Asantewaa-Tannor, P., Anornu, G.K., 2020.
use and population density on seasonal surface water quality using a modified
Spatial evaluation of land use variability on water quality of the Densu Basin, Ghana.
geographically weighted regression. Sci. Total Environ. 572, 450–466.
Water Sci. Technol. 20 (8), 3000.
Chen, T., Cui, D., Yuan, Y., Liu, J., Huang, M., Li, Y., 2021a. Changes in water quality and
Haith, D.A., 1976. Land use and water quality in New York rivers. J. Environ. Eng.
land use structure in the green-belt area around Shanghai. J. East China Normal
Division 102 (1), 1–15.
Univ. (Natural Science) 2021 (4), 81.
Hakamata, T., Hirata, T., Muraoka, K., 1992. Evaluation of land use and river water
Chen, Z., An, C., Tan, Q., Tian, X., Li, G., Zhou, Y., 2021b. Spatiotemporal analysis of
quality of the Tsukuba Mountains ecosystem, Japan. Catena (Amst) 19 (5), 427–439.
land use pattern and stream water quality in southern Alberta, Canada. J. Contam.
Harris, G.P., 2001. Biogeochemistry of nitrogen and phosphorus in Australian
Hydrol. 242, 103852.
catchments, rivers and estuaries: effects of land use and flow regulation and
Cheng, X., Huang, Y., Li, R., Pu, X., Huang, W., Yuan, X., 2020. Impacts of water
comparisons with global patterns. Marine Freshwater Res. 52 (1), 139–149.
temperature on phosphorus release of sediments under flowing overlying water.
Hoang, N.T., Kanemoto, K., 2021. Mapping the deforestation footprint of nations reveals
J. Contam. Hydrol. 235, 103717.
growing threat to tropical forests. Nature Ecol. Evol. 5 (6), 845–853.
Chiang, L.C., Wang, Y.C., Chen, Y.K., Liao, C.J., 2021. Quantification of land use/land
Horppila, J., 2019. Sediment nutrients, ecological status and restoration of lakes. Water.
cover impacts on stream water quality across Taiwan. J. Clean. Prod. 318, 128443.
Res. 160, 206–208.
Croft-White, M.V., Cvetkovic, M., Rokitnicki-Wojcik, D., Midwood, J.D., Grabas, G.P.,
International Food Policy Research Institute (IFPRI) and VEOLIA (2015) The murky future
2017. A shoreline divided: twelve-year water quality and land cover trends in Lake
of global water quality: new global study projects rapid deterioration in water quality. A
Ontario coastal wetlands. J. Great. Lakes. Res. 43 (6), 1005–1015.
White Paper. Washington, D.C. and Chicago, IL. Available at: https://www.ifpri.or
Dale, V.H., 1997. The relationship between land-use change and climate change. Ecol.
g/publication/murky-future-global-water-quality-new-global-study-projects-rapid-
Appl. 7 (3), 753–769.
deterioration-water (Assessed: 09 August 2023).
Dauer, D.M., Ranasinghe, J.A., Weisberg, S.B., 2000. Relationships between benthic
Jordan, S.J., Stoffer, J., Nestlerode, J.A., 2011. Wetlands as sinks for reactive nitrogen at
community condition, water quality, sediment quality, nutrient loads, and land use
continental and global scales: a meta-analysis. Ecosystems. 14, 144–155.
patterns in Chesapeake Bay. Estuaries 23 (1), 80–96.
Jung, K.W., Lee, S.W., Hwang, H.S., Jang, J.H., 2008. The effects of spatial variability of
Dawson, R.N., Murphy, K.L., 1972. The temperature dependency of biological
land use on stream water quality in a coastal watershed. Paddy Water Environ. 6,
denitrification. Water. Res. 6 (1), 71–83.
275–284.
Day Jr, J.W., Ko, J.Y., Rybczyk, J., Sabins, D., Bean, R., Berthelot, G., Twilley, R, 2004.
Kang, J.H., Lee, S.W., Cho, K.H., Ki, S.J., Cha, S.M., Kim, J.H., 2010. Linking land-use
The use of wetlands in the Mississippi Delta for wastewater assimilation: a review.
type and stream water quality using spatial data of fecal indicator bacteria and heavy
Ocean Coastal Manag. 47 (11-12), 671–691.
metals in the Yeongsan river basin. Water. Res. 44 (14), 4143–4157.
de Mello, K., Taniwaki, R.H., de Paula, F.R., Valente, R.A., Randhir, T.O., Macedo, D.R.,
Khan, M.N., Mohammad, F., 2014. Eutrophication: challenges and solutions.
Hughes, R.M., 2020. Multiscale land use impacts on water quality: Assessment,
Eutrophication: Causes, Consequences and Control: Volume 2, pp. 1–15.
planning, and future perspectives in Brazil. J. Environ. Manage 270, 110879.
Knee, K.L., Encalada, A.C., 2014. Land use and water quality in a rural cloud forest
Deeks, J. J., Higgins, J.P.T., Altman, D.G. (eds). (2022). Chapter 10: Analysing data and
region (Intag, Ecuador). River. Res. Appl. 30 (3), 385–401.
undertaking meta-analyses. In: Higgins, J.P.T., Thomas, J., Chandler, J., Cumpston,
Lai, Y.C., Tu, Y.T., Yang, C.P., Surampalli, R.Y., Kao, C.M., 2013. Development of a water
M., Li, T., Page, M.J., Welch, V.A. (eds). Cochrane Handbook for Systematic Reviews of
quality modeling system for river pollution index and suspended solid loading
Interventions version 6.3 (updated February 2022). Cochrane. Available from www.
evaluation. J. Hydrol. (Amst) 478, 89–101.
training.cochrane.org/handbook.

12
X. Shi et al. Water Research 260 (2024) 121946

Lee, S.W., Hwang, S.J., Lee, S.B., Hwang, H.S., Sung, H.C., 2009. Landscape ecological Ribolzi, O., Cuny, J., Sengsoulichanh, P., Mousquès, C., Soulileuth, B., Pierret, A.,
approach to the relationships of land use patterns in watersheds to water quality Sengtaheuanghoung, O., 2011. Land use and water quality along a Mekong tributary
characteristics. Landsc. Urban. Plan. 92 (2), 80–89. in Northern Lao PDR. Environ. Manage 47, 291–302.
Legendre, P., Legendre, L., 2012. Numerical Ecology [electronic resource], 3rd English ed. Rimba, A.B., Mohan, G., Chapagain, S.K., Arumansawang, A., Payus, C., Fukushi, K.,
Elsevier, Amsterdam; Boston. Avtar, R., 2021. Impact of population growth and land use and land cover (LULC)
Liaw, A., Wiener, M., 2002. Classification and regression by randomForest. R News 2 (3), changes on water quality in tourism-dependent economies using a geographically
18–22. weighted regression approach. Environ. Sci. Pollut. Res. 28, 25920–25938.
Liu, J., Niyogi, D., 2019. Meta-analysis of urbanization impact on rainfall modification. Rothenberger, M.B., Burkholder, J.M., Brownie, C., 2009. Long-term effects of changing
Sci. Rep. 9 (1), 7301. land use practices on surface water quality in a coastal river and lagoonal estuary.
Liu, J., Zhang, X., Wu, B., Pan, G., Xu, J., Wu, S., 2017. Spatial scale and seasonal Environ. Manage 44, 505–523.
dependence of land use impacts on riverine water quality in the Huai River basin, Rupinski, M.T., Dunlap, W.P., 1996. Approximating Pearson Product-Moment
China. Environ. Sci. Pollut. Res. 24, 20995–21010. Correlations from Kendall’s Tau and Spearman’s Rho. Educ. Psychol. Meas. 56 (3),
Lowrance, R., Altier, L.S., Newbold, J.D., Schnabel, R.R., Groffman, P.M., Denver, J.M., 419–429. https://doi.org/10.1177/0013164496056003004.
Robinson, J.L., 1997. Water quality functions of riparian forest buffers in Schepaschenko, D., Chave, J., Phillips, O.L., Lewis, S.L., Davies, S.J., Réjou-Méchain, M.,
Chesapeake Bay watersheds. Environ. Manage. Zo-Bi, I.C., 2019. The Forest Observation System, building a global reference dataset
Luo, Z., Shao, Q., Zuo, Q., Cui, Y., 2020. Impact of land use and urbanization on river for remote sensing of forest biomass. Sci. Data 6 (1), 198.
water quality and ecology in a dam dominated basin. J. Hydrol. (Amst) 584, 124655. Schwarzer, G., 2007. Meta: an R package for meta-analysis. R News 7 (3), 40–45.
Mansaray, A.S., Aamodt, J., Koroma, B.M., 2018. Water pollution laws in sierra leone—a Scully, M.E., 2010. Wind modulation of dissolved oxygen in Chesapeake Bay. Estuaries
review with examples from the UK and USA. Natural Res. 9 (11), 361–388. Coasts 33, 1164–1175.
Mao, D., Luo, L., Wang, Z., Wilson, M.C., Zeng, Y., Wu, B., Wu, J., 2018a. Conversions Simpson, I.M., Winston, R.J., Brooker, M.R., 2022. Effects of land use, climate, and
between natural wetlands and farmland in China: a multiscale geospatial analysis. imperviousness on urban stormwater quality: a meta-analysis. Sci. Total Environ.
Sci. Total Environ. 634, 550–560. 809, 152206.
Mao, D., Wang, Z., Wu, J., Wu, B., Zeng, Y., Song, K., Luo, L., 2018b. China’s wetlands Sliva, L., Williams, D.D., 2001. Buffer zone versus whole catchment approaches to
loss to urban expansion. Land Degrad. Develop. 29 (8), 2644–2657. studying land use impact on river water quality. Water. Res. 35 (14), 3462–3472.
Mehaffey, M.H., Nash, M.S., Wade, T.G., Ebert, D.W., Jones, K.B., Rager, A., 2005. Song, X.P., Hansen, M.C., Stehman, S.V., Potapov, P.V., Tyukavina, A., Vermote, E.F.,
Linking land cover and water quality in New York City’s water supply watersheds. Townshend, J.R., 2018. Global land change from 1982 to 2016. Nature 560 (7720),
Environ. Monit. Assess. 107, 29–44. 639–643.
Merghadi, A., et al., 2020. Machine learning methods for landslide susceptibility studies: Sun, L., Chen, J., Li, Q., Huang, D., 2020. Dramatic uneven urbanization of large cities
A comparative overview of algorithm performance. Earth. Sci. Rev. 207, 103225 throughout the world in recent decades. Nat. Commun. 11 (1), 5366.
https://doi.org/10.1016/j.earscirev.2020.103225. Sweeney, B.W., Bott, T.L., Jackson, J.K., Kaplan, L.A., Newbold, J.D., Standley, L.J.,
Miller, J.D., Hutchins, M., 2017. The impacts of urbanisation and climate change on Horwitz, R.J., 2004. Riparian deforestation, stream narrowing, and loss of stream
urban flooding and urban water quality: a review of the evidence concerning the ecosystem services. Proc. Nat. Acad. Sci. 101 (39), 14132–14137.
United Kingdom. J. Hydrol.: Reg. Stud. 12, 345–362. Tan, L., Ge, Z., Ji, Y., Lai, D.Y., Temmerman, S., Li, S., Tang, J., 2022. Land use and land
Ministry of Ecology and Environment (2019) HJ 355-2019: Technical specification for cover changes in coastal and inland wetlands cause soil carbon and nitrogen loss.
operation of wastewater on-line monitoring system (CODCr, NH3-N et al.). Available at: Global Ecol. Biogeogr. 31 (12), 2541–2563.
https://www.mee.gov.cn/ywgz/fgbz/bz/bzwb/jcffbz/201912/W020191 Tong, Y., Wang, M., Peñuelas, J., Liu, X., Paerl, H.W., Elser, J.J., Lin, Y., 2020.
227500068358506.pdf. (Accessed: 21 October 2023). Improvement in municipal wastewater treatment alters lake nitrogen to phosphorus
Morée, A.L., Beusen, A.H.W., Bouwman, A.F., Willems, W.J., 2013. Exploring global ratios in populated regions. Proc. Nat. Acad. Sci. 117 (21), 11566–11572.
nitrogen and phosphorus flows in urban wastes during the twentieth century. Global. Tram, V.N.Q., Somura, H., Moroizumi, T., Maeda, M., 2022. Effects of local land-use
Biogeochem. Cycles. 27 (3), 836–846. policies and anthropogenic activities on water quality in the upstream Sesan River
Morrissy, J.G., Currell, M.J., Reichman, S.M., Surapaneni, A., Megharaj, M., Crosbie, N. Basin, Vietnam. J. Hydrol.: Reg. Stud. 44, 101225.
D., Ball, A.S., 2021. Nitrogen contamination and bioremediation in groundwater and Tran, C.P., Bode, R.W., Smith, A.J., Kleppel, G.S., 2010. Land-use proximity as a basis for
the environment: a review. Earth. Sci. Rev. 222, 103816. assessing stream water quality in New York State (USA). Ecol. Indic. 10 (3),
Nafi’Shehab, Z., Jamil, N.R., Aris, A.Z., Shafie, N.S., 2021. Spatial variation impact of 727–733.
landscape patterns and land use on water quality across an urbanized watershed in Tromboni, F., Dilts, T.E., Null, S.E., Lohani, S., Ngor, P.B., Soum, S., Chandra, S., 2021.
Bentong, Malaysia. Ecol. Indicat. 122, 107254. Changing land use and population density are degrading water quality in the Lower
Nahm, F.S., 2016. Nonparametric statistical tests for the continuous data: the basic Mekong Basin. Water. (Basel) 13 (14), 1948.
concept and the practical use. Korean J. Anesthesiol. 69 (1), 8–14. Tu, J., 2009. Combined impact of climate and land use changes on streamflow and water
Nakagawa, S., Noble, D.W., Senior, A.M., Lagisz, M., 2017. Meta-evaluation of meta- quality in eastern Massachusetts, USA. J. Hydrol. (Amst) 379 (3-4), 268–283.
analysis: ten appraisal questions for biologists. BMC. Biol. 15 (1), 1–14. Tu, J., 2011. Spatial and temporal relationships between water quality and land use in
Olsson, L., Barbosa, H., Bhadwal, S., Cowie, A., Delusca, K., Flores-Renteria, D., northern Georgia, USA. J. Integr. Environ. Sci. 8 (3), 151–170.
Hermans, K., Jobbagy, E., Kurz, W., Li, D., Sonwa, D.J., Stringer, L., 2019. Land Tu, J., Xia, Z.G., 2008. Examining spatially varying relationships between land use and
Degradation. In: Climate Change and Land: an IPCC Special Report on Climate water quality using geographically weighted regression I: model design and
Change, Desertification, Land Degradation, Sustainable Land Management, Food evaluation. Sci. Total Environ. 407 (1), 358–378.
Security, and Greenhouse Gas Fluxes in Terrestrial Ecosystems. https://doi.org/ USGS (2019). How much water is there on Earth?. Available at:https://www.usgs.gov/spec
10.1017/9781009157988.006. ial-topics/water-science-school/science/how-much-water-there-earth (Assessed: 09
Pak, H.Y., Chuah, C.J., Yong, E.L., Snyder, S.A., 2021. Effects of land use configuration, August 2023).
seasonality and point source on water quality in a tropical watershed: a case study of Vaighan, A.A., Talebbeydokhti, N., Bavani, A.M., 2017. Assessing the impacts of climate
the Johor River Basin. Sci. Total Environ. 780, 146661. and land use change on streamflow, water quality and suspended sediment in the
Pastor, J., Solin, J., Bridgham, S.D., Updegraff, K., Harth, C., Weishampel, P., Dewey, B., Kor River Basin, Southwest of Iran. Environ. Earth. Sci. 76, 1–18.
2003. Global warming and the export of dissolved organic carbon from boreal van Vliet, M.T., Thorslund, J., Strokal, M., Hofstra, N., Flörke, M., Ehalt Macedo, H.,
peatlands. Oikos. 100 (2), 380–386. Mosley, L.M., 2023. Global river water quality under climate change and
Peiffer, S., Kappler, A., Haderlein, S.B., Schmidt, C., Byrne, J.M., Kleindienst, S., Planer- hydroclimatic extremes. Nature Rev. Earth Environ. 1–16.
Friedrich, B., 2021. A biogeochemical–hydrological framework for the role of redox- Verhoeven, J.T., Arheimer, B., Yin, C., Hefting, M.M., 2006. Regional and global
active compounds in aquatic systems. Nat. Geosci. 14 (5), 264–272. concerns over wetlands and water quality. Trends Ecol. Evol. 21 (2), 96–103.
Pielke Sr, R.A, 2005. Land use and climate change. Science (1979) 310 (5754), Viau, E.J., Goodwin, K.D., Yamahara, K.M., Layton, B.A., Sassoubre, L.M., Burns, S.L.,
1625–1626. Boehm, A.B., 2011. Bacterial pathogens in Hawaiian coastal streams—associations
Qiu, M., Wei, X., Hou, Y., Spencer, S.A., Hui, J., 2023. Forest cover, landscape patterns, with fecal indicators, land cover, and water quality. Water. Res. 45 (11), 3279–3290.
and water quality: a meta-analysis. Landsc. Ecol. 38 (4), 877–901. Vörösmarty, C.J., McIntyre, P.B., Gessner, M.O., Dudgeon, D., Prusevich, A., Green, P.,
Quinn, J.M., Stroud, M.J., 2002. Water quality and sediment and nutrient export from Davies, P.M., 2010. Global threats to human water security and river biodiversity.
New Zealand hill-land catchments of contrasting land use. New Zealand J. Marine Nature 467 (7315), 555–561.
Freshwater Res. 36 (2), 409–429. Vymazal, J., 2007. Removal of nutrients in various types of constructed wetlands. Sci.
Ramadas, M., Samantaray, A.K., 2018. Applications of remote sensing and GIS in water Total Environ. 380 (1-3), 48–65.
quality monitoring and remediation: a state-of-the-art review. Water Remed. 225- Wang, M., Bodirsky, B.L., Rijneveld, R., Beier, F., Bak, M.P., Batool, M., Strokal, M.,
246. 2024. A triple increase in global river basins with water scarcity due to future
Radwan, T.M., Blackburn, G.A., Whyatt, J.D., Atkinson, P.M., 2021. Global land cover pollution. Nat. Commun. 15 (1), 880.
trajectories and transitions. Sci. Rep. 11 (1), 12814. Wang, M., Janssen, A.B., Bazin, J., Strokal, M., Ma, L., Kroeze, C., 2022a. Accounting for
Records, R.M., Wohl, E., Arabi, M., 2016. Phosphorus in the river corridor. Earth. Sci. interactions between sustainable development goals is essential for water pollution
Rev. 158, 65–88. control in China. Nat. Commun. 13 (1), 730.
Regier, P.J., González-Pinzón, R., Van Horn, D.J., Reale, J.K., Nichols, J., Khandewal, A., Wang, W., Liu, C., Zhang, F., Tan, M.L., Shi, J., Zhang, Z., Xin, H., 2022b. Evaluation of
2020. Water quality impacts of urban and non-urban arid-land runoff on the Rio impacts of environmental factors and land use on seasonal surface water quality in
Grande. Sci. Total Environ. 729, 138443. arid and humid regions using structural equation models. Ecol. Indic. 144, 109546.
Ren, W., Zhong, Y., Meligrana, J., Anderson, B., Watt, W.E., Chen, J., Leung, H.L., 2003. Wei, W., Gao, Y., Huang, J., Gao, J., 2020. Exploring the effect of basin land degradation
Urbanization, land use, and water quality in Shanghai: 1947–1996. Environ. Int. 29 on lake and reservoir water quality in China. J. Clean. Prod. 268, 122249.
(5), 649–659. Wilkin, D.C., Jackson, R.W., 1984. Nonpoint water quality contributions from land use.
J. Environ. Syst. 13 (2), 127–136.

13
X. Shi et al. Water Research 260 (2024) 121946

Winkler, K., Fuchs, R., Rounsevell, M., Herold, M., 2021. Global land use changes are Yu, C., Xie, S., Song, Z., Xia, S., Åström, M.E., 2021. Biogeochemical cycling of iron
four times greater than previously estimated. Nat. Commun. 12 (1), 2501. (hydr-) oxides and its impact on organic carbon turnover in coastal wetlands: A
Winton, R.S., Teodoru, C.R., Calamita, E., Kleinschroth, F., Banda, K., Nyambe, I., global synthesis and perspective. Earth. Sci. Rev. 218, 103658.
Wehrli, B., 2021. Anthropogenic influences on Zambian water quality: hydropower Yu, D., Shi, P., Liu, Y., Xun, B., 2013. Detecting land use-water quality relationships from
and land-use change. Environ. Sci.: Process. Impacts 23 (7), 981–994. the viewpoint of ecological restoration in an urban area. Ecol. Eng. 53, 205–216.
Woldeab, B., Ambelu, A., Mereta, S.T., Beyene, A., 2019. Effect of watershed land use on Zalles, V., Hansen, M.C., Potapov, P.V., Parker, D., Stehman, S.V., Pickens, A.H.,
tributaries’ water quality in the east African Highland. Environ. Monit. Assess. 191, Kommareddy, I., 2021. Rapid expansion of human impact on natural land in South
1–13. America since 1985. Sci. Adv. 7 (14), eabg1620.
Woli, K.P., Nagumo, T., Kuramochi, K., Hatano, R., 2004. Evaluating river water quality Zhang, D.W., Li, Y.F., Sun, X., Zhang, F.S., Zhu, H.X., Liu, Y., Zhu, X.D., 2010.
through land use analysis and N budget approaches in livestock farming areas. Sci. Relationship between landscape pattern and river water quality in Wujingang
Total Environ. 329 (1-3), 61–74. region, Taihu Lake watershed. Environ. Sci. 31 (8), 1775–1783.
Wu, Q., Yue, K., Ma, Y., Heděnec, P., Cai, Y., Chen, J., Li, Y., 2022. Contrasting effects of Zhang, F., Chen, Y., Wang, W., Jim, C.Y., Zhang, Z., Tan, M.L., Rahman, H.A., 2022.
altered precipitation regimes on soil nitrogen cycling at the global scale. Glob. Impact of land-use/land-cover and landscape pattern on seasonal in-stream water
Chang. Biol. 28 (22), 6679–6695. quality in small watersheds. J. Clean. Prod. 357, 131907.
Xiao, R., Wang, G., Zhang, Q., Zhang, Z., 2016. Multi-scale analysis of relationship Zhang, J., Li, S., Dong, R., Jiang, C., Ni, M., 2019. Influences of land use metrics at multi-
between landscape pattern and urban river water quality in different seasons. Sci. spatial scales on seasonal water quality: a case study of river systems in the Three
Rep. 6 (1), 25250. Gorges Reservoir Area, China. J. Clean. Prod. 206, 76–85.
Xu, M., Zhang, T., Wang, D., Zhao, Y., Xie, Y., Ma, L., 2019. Review and prospect of water Zhang, M., Wei, X., 2021. Deforestation, forestation, and water supply. Science (1979)
pollution prevention and control of China in the forty years of reform and opening- 371 (6533), 990–991.
up. Chin. J. Environ. Manag 11, 65–71. Zhao, J., Lin, L., Yang, K., Liu, Q., Qian, G., 2015. Influences of land use on water quality
Xu, Q., Yan, T., Wang, C., Hua, L., Zhai, L., 2023. Managing landscape patterns at the in a reticular river network area: A case study in Shanghai, China. Landsc. Urban.
riparian zone and sub-basin scale is equally important for water quality protection. Plan. 137, 20–29.
Water. Res. 229, 119280. Zhu, L., Chen, Y., Wang, Y., Wei, Y., Zheng, H., Zhang, Y., 2023. A comprehensive
Yadav, S., Babel, M.S., Shrestha, S., Deb, P., 2019. Land use impact on the water quality analysis of impacts of socio-economic development and land use on river water
of large tropical river: Mun River Basin, Thailand. Environ. Monit. Assess. 191, 1–22. quality in a megacity-region: a case study. Environ. Res. Commun. 5 (2), 025006.
Yang, Y., Roderick, M.L., Guo, H., Miralles, D.G., Zhang, L., Fatichi, S., Yang, D., 2023. Zou, K.H., Tuncali, K., Silverman, S.G., 2003. Correlation and simple linear regression.
Evapotranspiration on a greening Earth. Nat. Rev. Earth Environ. https://doi.org/ Radiology. 227 (3), 617–628.
10.1038/s43017-023-00464-3.

14