https://orcid.org/0000-0002-4208-1625
Abstract
Investigating the influence of land use change on climate elasticity of water quality (CEWQ) at multiple spatial scales is very critical for sustainable water management policies. The current research work investigated land use change–water–climate nexus through the lens of 32 water quality monitoring sites located at major rivers of Pakistan. The novelty of the current research work is the assessment of the instability consequences of precipitation and temperature CEWQ indicators owing to land use dynamics (2001–2019) at both sub-watershed and buffer scales. Precipitation elasticity values are lower and spatially homogeneous in comparison with temperature elasticity. Majority CEWQ-land use correlation coefficients have not shown any temporal trend with land use change except a few CEWQ indicators, namely pH, CO3, F, Ca, SiO2, silt, and clay. Temperature CEWQ developed many linear models with land use in comparison with precipitation CEWQ. A small number of land use classes cause instability consequences at the buffer scale in comparison with the sub-watershed scale. Savanna, shrublands, and ice and snow decline instability consequences of CEWQ indicators at both spatial scales. The identified land use classes that bring stability in CEWQ indicators are recommended to be incorporated in watershed management policies to bring sustainability in the aquatic environment.
HIGHLIGHTS
Elasticity technique was used to gauge the response of water quality parameters to climatic factors.
Water–land–climate nexus was developed to highlight interconnections among water, land, and climate.
CEWQ showed instability consequences owing land use change at the buffer and sub-watershed scales.
Precipitation CEWQ is homogenous in space.
Sediments’ responses to climatic drivers are compact and lower.
INTRODUCTION
Land use dynamics cause environmental changes on a global scale that deteriorate surface water quality. Climatic systems and surface water are interlinked with each other, so variation in one system impacts the other (Cantonati et al. 2020). Climate change has a profound influence on surface water quantity and quality. Climatic and nonclimatic determinants impair surface water quality (Panthi et al. 2017). Water quality deterioration affects the functionality of water systems, namely drinking water systems, irrigation, recreation, etc. (Meneses et al. 2010). It is of utmost importance to examine the response of water quality indicators to climate drivers and their instability consequences owing to land use change at the sub-watershed and buffer scales.
Literature shows that climatic determinants, namely precipitation and air temperature, impair surface water quality (Kundzewicz & Krysanova 2010). The variations in climatic drivers cause negative consequences in the environment for water quality and quantity (Akpodiogaga-a & Odjugo 2010). Global warming enhances air temperature, which in turn amplifies water temperature and speeds up chemical processes in surface waters (Woolway & Merchant 2017). Variation in precipitation owing to climate change influences river flow, which impacts the transport and dilution of contaminants. It is of the utmost importance to gauge the response of water quality indicators to varying precipitation and air temperature to formulate water policies accordingly.
Land use is an important terrestrial determinant that should be incorporated into land development policy formulation (Afed Ullah et al. 2018). The impacts of land dynamics on the water environment are extensively studied by developing associations between water quality indicators and land use (Tong et al. 2012). Natural land conversion into agriculture and urban areas to fulfill the growing demand for food and accommodation has threatened the ecosystem's health by enhancing surface runoff and pollution load (Lee et al. 2009; Shah et al. 2020). This is the cry of the day to assess the land development impacts on surface water quality to formulate water policy accordingly (Karr & Dudley 1981).
Generally, water quality and land use associations are developed at two spatial scales, namely, parallel buffer and watershed scale, which give an idea of the nearest and furthest pollution emitting sources’ influence on surface water quality (Afed Ullah et al. 2018). The size of parallel buffer ranges from 100 to 1,500 m (White 1995; Khan et al. 2021). This approach helps us to judge how the land use-surface water quality relationship changes with increasing distance from the water body. It identifies hotspots for land management at a particular spatial scale (Jaiswal & Pandey 2019). Land use-water quality relationships are extensively developed at multiple spatial scales while land use-CEWQ relationships have not been practiced at multiple spatial scales, which motivated the author to investigate.
CEWQ is a nonparametric approach that gauges the response of water quality indicators to climatic drivers (Jiang et al. 2014). Literature shows that land use-CEWQ and land use-soil relationships are not stable in time and space. Precipitation elasticity is highly sensitive to socio-economic and topographic factors in comparison with temperature elasticity on a global scale. CEWQ is highly unstable to land use change in comparison with surficial geology. Watersheds with a higher ratio of forest, urban, and agricultural land are more prone to the instability consequences of CEWQ in comparison with ran. Watershed health is inversely proportional to intense anthropogenic and agricultural activities (Ervinia et al. 2019; Jiang et al. 2019). The land use-CEWQ relationship is assessed on one spatial (watershed scale) and one temporal (one-year land use only) scale. To fill the gap, the author investigated the instability consequences of CEWQ at multiple spatial (sub-watershed and parallel buffer) and temporal scales (2001, 2005, 2010, 2015, and 2019 land use). To examine a water–land use change–climate nexus, the following questions are important:
Which spatial scale causes instability consequences in CEWQ?
What is the temporal trend (increasing or decreasing) of land use-CEWQ linkage?
The principal aim of the current research work is to assess the instability consequences of CEWQ at multiple spatial (sub-watershed and parallel buffer) and temporal scales (2001, 2005, 2010, 2015, and 2019 land use). First, this research work aims to check the homogeneity of precipitation and temperature elasticity in space. Second, it aims to check the instability consequences of CEWQ at multiple spatial (sub-watershed and parallel buffer) and temporal scales (2001, 2005, 2010, 2015, and 2019 land use). To achieve the aforementioned goals, 32 water quality monitoring stations were selected as a case study, which are mainly located in the upper northern area of Pakistan.
STUDY AREA AND METHODS
Watershed description
Study area map demonstrating monitoring sites, DEM, demarcated sub-watersheds and parallel buffer.
Study area map demonstrating monitoring sites, DEM, demarcated sub-watersheds and parallel buffer.
The Indus is the longest river in Pakistan and passes through the entire length of the country. It originates in Tibet, China, enters Kashmir, and flows through Pakistan from the extreme north to the south before it finally falls into the Arabian Sea near Karachi. This is the longest river in Pakistan, whose drainage area and annual flow are approximately 1,165,000 km2 and 243 km3, respectively. The Jhelum River lies between India and Pakistan, with a length and discharge of 725 km and 221.9 m3/s, respectively. The river passes through the Kashmir valley and Punjab and discharges into the Indus River. The Chenab River flows through India and Pakistan, having a length of 960 km. The river enters the plains of Punjab from the Jammu and Kashmir region. The Ravi River is a transboundary river between India and Pakistan, having a length of 720 km. As per the Indus Water Treaty, the Ravi River water is allocated to India. The Sutlej River passes through India and Pakistan, having a length and basin size of 1,450 km and 395,000 km2, respectively. The Kabul River originates from Afghanistan and enters Pakistan and has a length of 700 km.
Data collection
Digital elevation model and terrestrial determinants data
The digital elevation model (DEM) was used to delineate sub-watersheds as demonstrated in Figure 1. DEM data were obtained from the Shuttle Radar Topography Mission (SRTM; https://portal.opentopography.org/raster?opentopoID=OTSRTM.082015.4326.1). Land use data (2001, 2005, 2010, 2015, and 2019) were obtained from Landsat (https://www.usgs.gov/core-science-systems/nli/landsat/landsat-data-access?qt-science_support_page_related_con=0#qt-science_support_page_related_con).
Water quality data
Water quality data were collected from the global water quality data portal (https://gemstat.org/data/data-portal/) and Pakistan Water & Power Development Authority (WAPDA; http://www.wapda.gov.pk/). This study is based on sediments and physio-chemical water quality parameters which include Ca, Mg, Na, K, CO3, HCO3, Cl, SO4, NO3, F, cations, anions, SiO2, Fe, B, DS by evaporation, EC, pH, Res CO3 me/l, SAR, Temp, PPM, sand, silt, and clay (Table 1).
Summary of the data
| Stations | Historical record of water quality data (Years) | Number of water quality parameters | |
|---|---|---|---|
| 1 | Neelum River at Muzaffarabad | 1963–2008 | Ca, Mg, Na, K, CO3, HCO3, Cl, SO4, NO3, F, cations, anions, SiO2, Fe, B, DS by evaporation, EC, pH, Res CO3 me/l, SAR, Temp, PPM, sand, silt, and clay |
| 2 | Golan River at Babuka | ||
| 3 | Neelam River at Nosheri | ||
| 4 | Jhelum River at Dolai | ||
| 5 | Shigar River at Shigar | ||
| 6 | Gilgit River at Gilgit | ||
| 7 | Soan River at Chirah | ||
| 8 | Soan River at Rawalpindi | ||
| 9 | Tochi River at Tangi Post | ||
| 10 | Tank Zam River at Jandola | ||
| 11 | Sill River near Chahan | ||
| 12 | Khost River at Chappar Rift | ||
| 13 | Hunza River at Danyore Bridge | ||
| 14 | Gilgit River at Alam Bridge | ||
| 15 | Beji River at Babar Kach | ||
| 16 | Jhelum River at Mangla | ||
| 17 | Garban River at Karora | ||
| 18 | Hub River at Karpus Anivat | ||
| 19 | Indus River at Kachora | ||
| 20 | Shyok River at Yoga | ||
| 21 | Indus River at Dadomoro | ||
| 22 | Indus River at Khairabad | ||
| 23 | Swat River near Kalam | ||
| 24 | Swat River near Chakdara | ||
| 25 | Bara River at Jhansi Post | ||
| 26 | Kabul River at Nowshera | ||
| 27 | Kunhar River at Talhata | ||
| 28 | Gomal River at Kot Murtaza | ||
| 29 | Poonch River at Kotli | ||
| 30 | Chitral River at Chitral | ||
| 31 | Soan River at Dhok Pathan | ||
| 32 | Daraban River at Zam Tower | ||
| Stations | Historical record of water quality data (Years) | Number of water quality parameters | |
|---|---|---|---|
| 1 | Neelum River at Muzaffarabad | 1963–2008 | Ca, Mg, Na, K, CO3, HCO3, Cl, SO4, NO3, F, cations, anions, SiO2, Fe, B, DS by evaporation, EC, pH, Res CO3 me/l, SAR, Temp, PPM, sand, silt, and clay |
| 2 | Golan River at Babuka | ||
| 3 | Neelam River at Nosheri | ||
| 4 | Jhelum River at Dolai | ||
| 5 | Shigar River at Shigar | ||
| 6 | Gilgit River at Gilgit | ||
| 7 | Soan River at Chirah | ||
| 8 | Soan River at Rawalpindi | ||
| 9 | Tochi River at Tangi Post | ||
| 10 | Tank Zam River at Jandola | ||
| 11 | Sill River near Chahan | ||
| 12 | Khost River at Chappar Rift | ||
| 13 | Hunza River at Danyore Bridge | ||
| 14 | Gilgit River at Alam Bridge | ||
| 15 | Beji River at Babar Kach | ||
| 16 | Jhelum River at Mangla | ||
| 17 | Garban River at Karora | ||
| 18 | Hub River at Karpus Anivat | ||
| 19 | Indus River at Kachora | ||
| 20 | Shyok River at Yoga | ||
| 21 | Indus River at Dadomoro | ||
| 22 | Indus River at Khairabad | ||
| 23 | Swat River near Kalam | ||
| 24 | Swat River near Chakdara | ||
| 25 | Bara River at Jhansi Post | ||
| 26 | Kabul River at Nowshera | ||
| 27 | Kunhar River at Talhata | ||
| 28 | Gomal River at Kot Murtaza | ||
| 29 | Poonch River at Kotli | ||
| 30 | Chitral River at Chitral | ||
| 31 | Soan River at Dhok Pathan | ||
| 32 | Daraban River at Zam Tower | ||
Methods
Spatial scale analysis
Sub-watersheds were delineated for each outlet using the Hydrology tool of ArcMap 10.2. The upper sub-watershed contributes nutrient load to each water quality monitoring site. Similarly, a parallel buffer of 1,000 m was demarcated around each waterway located in each sub-watershed. Land use data for 2001, 2005, 2010, 2015, and 2019 were extracted for both spatial scales, i.e., sub-watershed and parallel buffer. The extracted land use data were linked with CEWQ parameters. The above-mentioned technique will enable us to see how the temporal land use-CEWQ relationship varies as distance increases from the monitoring site. It will highlight land use conservative efforts at a specific spatial scale.
Statistical technique
Pearson's correlation and stepwise multiple linear regressions were carried out to make linkages between land use and CEWQ at multiple spatial scales using 2001, 2005, 2010, 2015, and 2019 land use data. The main aim of the statistical analysis was to identify land use variables that have the strongest association with CEWQ parameters. The multiyear land use-CEWQ correlation coefficient will highlight the increasing or decreasing pattern at specific spatial scales for the CEWQ parameter. Box plot analysis was carried out to check the spatial homogeneity of precipitation and temperature elasticity. The one having a compact distribution will show a homogeneous response in space.
Nonparametric estimators of the CEWQ
(P, WQ) depends on two variables value, i.e., WQ and P.
and 
represent mean annual precipitation and water quality, respectively. Robust precipitation elasticity estimator is demonstrated by Equation (3):
is the mean annual water quality, 
is the mean annual air temperature, and 
is the mean annual precipitation.Similarly, 
is the water quality at any instant t, 
is the air temperature at any instant t, and 
is the precipitation at any instant t.
RESULTS AND DISCUSSION
Impact of land use change on CEWQ at multiple spatial scales
Extraction of land use data at the sub-watershed scale using GIS
Percentage land use composition of the delineated sub-watersheds.
Figure 2 shows the results of the sub-basin scale land use data extraction. As there were multiple stations, we only compared the last station (Hunza River at Danyore Bridge) for each year's land use data. The figure makes it evident that savannas cover 0–18% of the land, water covers 18–39% of the land, and evergreen needle forests cover the remaining 9–100% of the area.
Similarly, according to statistics on land use from 2005, we can observe that savanna land use has grown by 1% or 0–19%, water land use has changed by 19–41%, and evergreen needle leaf land use has increased by 41–100%.
Savanna land use in 2010 remained unchanged at 0–19%; water land use changed to 19–42%, and evergreen needle leaf land use increased to 42–100%.
In 2015, the land usage for savannas remained the same at 0–19%, but the land use for water changed to 19–44%, and the land use for evergreen needle leaves increased to 44–100%.
According to the 2019 land use data, the land use for savannas has fallen by 1%, or 0–18%, while the land use for water has changed by 18–42% and the land use for evergreen needle leaves has increased by 42–100%.
Extraction of land use at the parallel buffer scale using GIS
Percentage land use composition of the demarcated parallel buffer.
Figure 3 shows the results of the parallel buffer scale land use data extraction. According to data on land use from 2001, savanna land usage is between 0 and 16%, water land use is between 16 and 39%, and evergreen needle leaf land use is between 39 and 100%. Similarly, according to statistics on land use from 2005, we can see that savanna land use has increased by 1%, or 0–17%, water land use has changed by 17–42%, and evergreen needle leaf land use has increased by 42–100%. The land use for savannas is 0–17%, the land use for water is change that is 17–41%, and the land use for evergreen needle leaf is 41–100%, according to data on land use from 2010. According to data on land use from 2015, we can observe that savanna land usage is 0–17%, water land use is 17–43%, and evergreen needle leaf land use is 43–100%. While the land use for savannas is up 1% in 2019 (from 0 to 18%), the land use for water is also changing (from 17 to 45%), and the land use for evergreen needle leaf is from 45 to 100%.
Spatial homogeneity of CEWQ
Box plot for precipitation and temperature elasticity.
εP results are superior to εT due to spatial homogeneous response. Moreover, the results of εP are lower than εT. εP results are physically possible in comparison with εT. The results implied that εP is superior to εT (Afed Ullah et al. 2018). The lower values of εP may be due to dilution effects because surface runoff sweeps all kinds of pollutants from the land surface to nearby waterbodies. In comparison, εT is complex owing to complex biogeochemical processes (Khan et al. 2017a, 2017b). An increase in air temperature due to global warming can cause variations in surface water quality (Hosseini et al. 2017; Pompei et al. 2020).
Correlation between CEWQ and land use change
A Pearson's correlation analysis was conducted to develop a relationship between CEWQ and land use at the buffer and sub-watershed scales.
Sensitivity of precipitation elasticity to land use changes at the buffer scale
Pearson's correlation coefficient between precipitation CEWQ indicators and land use at the buffer Scale.
Pearson's correlation coefficient between precipitation CEWQ indicators and land use at the buffer Scale.
Sensitivity of precipitation elasticity to land use changes at the sub-watershed scale
Pearson's correlation coefficient between precipitation CEWQ indicators and land use at the sub-watershed scale.
Pearson's correlation coefficient between precipitation CEWQ indicators and land use at the sub-watershed scale.
Sensitivity of air temperature elasticity to land use changes at the buffer scale
Pearson's correlation coefficient between temperature CEWQ indicators and land use at the buffer scale.
Pearson's correlation coefficient between temperature CEWQ indicators and land use at the buffer scale.
Sensitivity of air temperature elasticity to land use changes at the sub-watershed scale
Pearson's correlation coefficient between temperature CEWQ indicators and land use at the sub-watershed scale.
Pearson's correlation coefficient between temperature CEWQ indicators and land use at the sub-watershed scale.
Physical insights to the correlation results
Negative elasticity and negative correlation demonstrate that under the same air temperature and same precipitation conditions, those buffers and sub-watersheds tend to reduce the instability consequences of CEWQ (decline pollutant loads and concentrations) and vice versa. The temporal declining trend of the CEWQ-land use correlation coefficient (negative correlation) means that the strength of a relationship is decreasing, which declines the instability consequences of CEWQ (decline pollutant loads and concentrations) at that particular buffer and sub-watershed scale and vice versa.
Savanna-εP (P, F) correlation coefficient (negative) exhibited an increasing temporal trend at the sub-watershed scale from 2001 to 2019. Deciduous broadleaf forest-εP (P, pH) correlation coefficient (positive) showed a decreasing trend at the buffer scale, while savanna-εP (P, CO3) correlation coefficient (positive) exhibited an increasing temporal trend at the buffer scale from 2001 to 2019.
Savanna-εP (P, F) correlation coefficient (negative) exhibited an increasing temporal trend at the sub-watershed scale from 2001 to 2019. Savanna-εT (T, Ca) and cropland-εT (T, SiO2) correlation coefficient (negative) exhibited an increasing temporal trend at the sub-watershed scale, while savanna-εT (T, silt) and savanna-εT (T, clay) correlation coefficient (negative) showed a temporal increasing trend with savanna land use at the buffer scale from 2001 to 2019. The above discussion shows that savanna, shrublands, and ice and snow cover play a friendly role in stabilizing CEWQ. The identified land use classes which bring stability in CEWQ indicators are recommended to be incorporated into watershed management policies to bring sustainability to the surface water environment.
Linear models between CEWQ and land use change
To quantitatively investigate the association between land use change and CEWQ, stepwise regression was carried out at two spatial scales, namely sub-watershed and buffer scales, as demonstrated in Table 2. Collectively, it is evident that the temperature elasticity of water quality has developed many linear empirical equations with land use in comparison with precipitation elasticity of water quality. CEWQ mainly builds relationships with forest, water, cropland, natural vegetation mosaic, urban, and grassland.
Stepwise multiple linear regression models for precipitation and temperature elasticity at the buffer and sub-watershed scales
| Elasticity, scale, year | CEWQ | Forest | Water | UR | GL | CN | R2 |
|---|---|---|---|---|---|---|---|
| εP, buffer, 2005 | K | + | 0.994 | ||||
| εP, buffer, 2019 | pH | + | 0.955 | ||||
| εT, buffer, 2005 | Cation | − | 0.881 | ||||
| pH | + | + | − | 0.983 | |||
| SO4 | + | + | − | 0.94 | |||
| SAR | + | 0.958 | |||||
| εT, buffer, 2010 | SiO2 | + | 0.724 | ||||
| pH | + | 0.616 | |||||
| εT, buffer, 2015 | Anion | + | 0.881 | ||||
| pH | + | 0.981 | |||||
| SAR | + | 0.602 | |||||
| εT, buffer, 2019 | K | + | 0.995 | ||||
| SO4 | + | + | 0.899 | ||||
| pH | + | 0.866 | |||||
| SAR | + | 0.836 | |||||
| Sand | 0.852 | ||||||
| εP, watershed, 2001 | pH | + | 0.695 | ||||
| εP, watershed, 2005 | HCO3 | + | 0.540 | ||||
| εP, watershed, 2010 | K | − | 0.639 | ||||
| εP, watershed, 2015 | K | + | 0.998 | ||||
| εP, watershed, 2019 | Mg | + | 0.628 | ||||
| εT, watershed, 2001 | K | + | 0.755 | ||||
| Ca | + | 0.734 | |||||
| εT, watershed, 2005 | DS | + | 0.848 | ||||
| pH | + | 0.896 | |||||
| εT, watershed, 2010 | Ca | − | 0.713 | ||||
| HCO3 | − | − | 0.706 | ||||
| Sand | 0.893 | ||||||
| εT, watershed, 2015 | Cl | + | 0.541 | ||||
| εT, watershed, 2019 | NO3 | + | 0.907 | ||||
| F | + | − | 0.976 | ||||
| pH | 0.616 |
| Elasticity, scale, year | CEWQ | Forest | Water | UR | GL | CN | R2 |
|---|---|---|---|---|---|---|---|
| εP, buffer, 2005 | K | + | 0.994 | ||||
| εP, buffer, 2019 | pH | + | 0.955 | ||||
| εT, buffer, 2005 | Cation | − | 0.881 | ||||
| pH | + | + | − | 0.983 | |||
| SO4 | + | + | − | 0.94 | |||
| SAR | + | 0.958 | |||||
| εT, buffer, 2010 | SiO2 | + | 0.724 | ||||
| pH | + | 0.616 | |||||
| εT, buffer, 2015 | Anion | + | 0.881 | ||||
| pH | + | 0.981 | |||||
| SAR | + | 0.602 | |||||
| εT, buffer, 2019 | K | + | 0.995 | ||||
| SO4 | + | + | 0.899 | ||||
| pH | + | 0.866 | |||||
| SAR | + | 0.836 | |||||
| Sand | 0.852 | ||||||
| εP, watershed, 2001 | pH | + | 0.695 | ||||
| εP, watershed, 2005 | HCO3 | + | 0.540 | ||||
| εP, watershed, 2010 | K | − | 0.639 | ||||
| εP, watershed, 2015 | K | + | 0.998 | ||||
| εP, watershed, 2019 | Mg | + | 0.628 | ||||
| εT, watershed, 2001 | K | + | 0.755 | ||||
| Ca | + | 0.734 | |||||
| εT, watershed, 2005 | DS | + | 0.848 | ||||
| pH | + | 0.896 | |||||
| εT, watershed, 2010 | Ca | − | 0.713 | ||||
| HCO3 | − | − | 0.706 | ||||
| Sand | 0.893 | ||||||
| εT, watershed, 2015 | Cl | + | 0.541 | ||||
| εT, watershed, 2019 | NO3 | + | 0.907 | ||||
| F | + | − | 0.976 | ||||
| pH | 0.616 |
UR, urban and built-up; GL, grasslands; CN, cropland/natural vegetation mosaic.
At the buffer scale, precipitation CEWQ developed an empirical equation with forest and water while temperature CEWQ developed a relationship with the forest, water, and crop/natural vegetation mosaic. Temperature CEWQ developed many linear empirical equations with land use in comparison with precipitation CEWQ (Khan et al. 2017a, 2017b).
At the sub-watershed scale, for precipitation, CEWQ developed an empirical equation with forest, water, grassland, and urban, while temperature CEWQ developed a relationship with the forest, water, grassland, cropland, natural vegetation mosaic, and urban areas. Temperature CEWQ developed many linear empirical equations with land use in comparison with precipitation CEWQ (Khan et al. 2017a, 2017b).
The above discussion shows that a small number of land uses causes instability consequences at the buffer scale in comparison with the sub-watershed scale. Temperature CEWQ developed many linear models with various kinds of land uses that may be attributed to complex biogeochemical processes occurring at various land use levels (Khan et al. 2017a, 2017b). An increase in air temperature as a consequence of global warming can cause variation in surface water quality (Hosseini et al. 2017).
ECOLOGICAL AND MANAGEMENT IMPLICATIONS
The current study reveals that land use dynamics cause instability consequences in CEWQ. The developed linear models between land use change and CEWQ can be used in formulating water policy to enhance stream health at buffer and sub-watershed scales. CEWQ is a handy approach to associating climatic variables, land use, and water quality parameters.
CEWQ developed many useful linear models with land use that vary with land use dynamics at both spatial scales. Watershed managers can take guidance from the developed CEWQ-land use linear models to bring stability to the water environment via best management practices (BMPs) at buffer and sub-watershed scales. The present study shows that various land uses cause instability consequences in CEWQ differently, which vary with time and scale. Watershed managers should consider the ill effects of land use dynamics before developing the land. To reduce the instability consequences of CEWQ, land development policy should be strictly followed to bring sustainability to the water environment.
The developed empirical linear models can be used in predicting the influence of land use on CEWQ at the buffer and sub-watershed scales. These models can be used in formulating water policy to adopt preventive measures to reduce nonpoint source pollution originating from various land uses. The use of the proposed models will help in controlling sediments and diffuse pollution in the study area. The developed linear models, in combination with GIS, will help reduce the instability consequences of CEWQ, which will help integrate natural ecosystems with dense settlements.
FUTURE WORK
The present study was designed to identify land use classes that cause instability consequences of CEWQ. The results of the current study will be useful for climate change experts, watershed managers, and regional land development policymakers. In the future, multiple parallel and circular buffers will be incorporated to scrutinize the key land use and land cover classes (edge density, patch density, etc.) causing instability consequences at multiple spatial scales.
CONCLUSION
Understanding the potential impacts of climatic drivers on water quality and its instability consequences owing to land use dynamics at multiple spatial scales is critical for sustainable watershed management. This research was conducted to study the effect of land use change on CEWQ at multiple spatial scales, i.e., buffer and sub-watershed scales. The current study is based on 32 water quality monitoring stations located in major rivers of Pakistan that were selected as the study area.
Precipitation CEWQ values are lower and more spatially homogeneous in comparison with temperature CEWQ values. Temperature CEWQ indicators have many linear empirical models in comparison with precipitation CEWQ indicators. The majority of precipitation CEWQ parameters did not show any specific temporal increasing or decreasing trends with land use change. Temperature CEWQ developed many linear models with land use in comparison with precipitation CEWQ. Savanna, shrublands, and ice and snow decline instability consequences of CEWQ indicators at both spatial scales.
Few CEWQ indicators showed well-defined temporal increasing or decreasing trends with land use change, namely pH, CO3, F, Ca, SiO2, silt, and clay. Savanna-εP (P, F) correlation coefficient (negative) exhibited an increasing temporal trend at the sub-watershed scale from 2001 to 2019. Deciduous broadleaf forest-εP (P, pH) correlation coefficient (positive) showed a decreasing trend at the buffer scale, while savanna-εP (P, CO3) correlation coefficient (positive) exhibited an increasing temporal trend at the buffer scale from 2001 to 2019. Savanna-εP (P, F) correlation coefficient (negative) exhibited an increasing temporal trend at the sub-watershed scale from 2001 to 2019.
Savanna-εT (T, Ca) and cropland-εT (T, SiO2) correlation coefficient (negative) exhibited an increasing temporal trend at the sub-watershed scale, while savanna-εT (T, silt) and savanna-εT (T, clay) correlation coefficient (negative) showed a temporal increasing trend with savanna land use at the buffer scale from 2001 to 2019. The present study identified land use classes (savanna, shrublands, and ice and snow cover) that bring stability to CEWQ indicators and is recommended to be incorporated into watershed management policies to bring sustainability to the surface water environment.
Results of the current study provide useful information regarding water quality indicators’ sensitivity to climatic drivers and their instability consequences owing to land use dynamics in multiple watersheds in the upper part of Pakistan. Empirical models can be useful to highlight key indicators to improve stream water quality. Additional research is required to apportion watershed pollution sources (biogeochemical hot spots).
DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.
CONFLICT OF INTEREST
The authors declare there is no conflict.
%20cropped.png?versionId=7149)