Impact of climate change on spatiotemporal variations in the vegetation cover and hydrology of district Nowshera Open Access

As a crucial component of ecosystems and an indication of climate change and anthropogenic activity, vegetation can connect soil, atmosphere, and water (Vereecken et al. 2010). Biogeochemical cycles in ecosystems are largely governed by vegetation cover (Gerten et al. 2004; Troch et al. 2009), however they are disrupted by human activities. Natural disturbances are caused by changes in vegetation, which alter biosphere–atmosphere interactions (Liang et al. 2015). Disruptions in the hydrological cycle can readily cause severe erosion (Ludwig et al. 2005), resulting in soil deterioration, reduced land productivity, and degraded lakes, streams, and estuaries (Pimentel & Kounang 1998; Marques et al. 2008). Climate and vegetation have a tight relationship, thus changes in one affect changes in the other. Changes in vegetation cover can have an impact on the climate system by changing the components of an area's climate. Similarly, changes in the climate influence changes in the dynamics of vegetation. Because of long-term interactions between plant and environment, vegetation responds to regional and global climate change through changes in spatial distribution and temporal phenology. These intricate relationships between plant physiological processes and climate variability can be described by dynamic ecosystems and their surroundings (Sitch et al. 2003; Sarker et al. 2019).

Remote sensing has made a significant contribution to the monitoring of vegetation and land cover change (Atif et al. 2015; Sarker 2020, 2021). The Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Water Index (NDWI) are key indices derived using satellite data for time-series analysis of vegetation cover change and hydrological events. Remote sensing and Geographic Information System (GIS) play important roles in time-series analysis of climate change influence on vegetation and land cover change caused by natural catastrophes such as floods. The advent of GIS, along with the synoptic view and repeating nature of remote satellite sensing data, has resulted in improved mapping and real-time monitoring of vegetation, urbanization, and flood inundation (Smith 1997; Sanyal & Lu 2004). The NDVI is a very accurate indicator for studying the spatiotemporal change in the vegetation cover of a region (Gao et al. 2019), whereas the NDWI is employed to evaluate the influence of severe rainfall and floods on land cover change. Mann Kendel and linear regression models are used to examine the trend of vegetation cover and land cover variation (Jiang et al. 2015).

Since the 1920s, Pakistan has been badly impacted by floods during the monsoon season of July–September. Since its inception, it has witnessed heavy floods in 1955, 1973, 1976, 1980, 1988, 1992 (Solheim et al. 2001), 2010, 2014, 2015, 2016, 2017, 2019, 2020, and 2021 (EM-DAT 2021). According to the global climate risk index 2020 (1999–2018) and 2021 (2000–2019), Pakistan is ranked fifth and eighth most vulnerable to extreme weather conditions caused by climate change, respectively (Eckstein et al. 2019, 2021), reflecting its persistent position in the top ten most defenseless countries on the planet. Majority of Pakistan has an arid to semi-arid climate with an average annual rainfall of 250 mm, whereas the north has a humid and highland environment due to high elevations (Khan 2003). Moving from south to north, the aridity gradually lessens as the rainfall continuously increases (Shams 2006). The country's famous two annual rainy spells, the winter depressions (December to March) and the summer monsoon (July to September), are interrupted by brief periods of light rain caused by local thunderstorms and convections. The annual precipitation at the Himalayas of Pakistan ranges from 760 to 1,270 mm which provides 72% of the mean annual flow of the Indus River (Tariq & Van De Giesen 2012). The river is sediment-laden, causing run-off and catastrophic flooding as a result of human intrusion through land removal, land use change, embankments, and dam and barrage building (Sinha 2009). The 2010 flooding disaster in the Indus-Kabul river system killed 1,961 persons and caused infrastructure losses totaling 9,500,000 USD (EM-DAT 2021). In terms of average annual flow (7,610 m³), length (3,180 km), and drainage area, the Indus River is one of the world's largest rivers (960,000 km²). Out of the 960,000 km² drainage areas, approximately 453,247 km² is in the mountains and foothills, while the rest (506,753 km²) is in Pakistan's semi-arid region (Khan et al. 2011). The river is split into two sections: upper and lower Indus. The upper Indus runs from the Tibetan Plateau to Mount Kailas in the northern Himalayas, with some flow from the Karakoram to the Guddu barrage, while the lower Indus runs from the Guddu barrage to the Arabian Sea. With an annual average sediment load of 291 million tons, the river is one of the most sediment-carrying rivers on the planet. The sediment load has impaired the water quality, which has been exacerbated by human activities. Many areas have observed rapid siltation, such as a rate of 200 million tons/year at Tarbella dam (Sloff 1997).

Khyber Pakhtunkhwa (KP) has many prominent rivers, including the Indus, Swat, Kabul, Kunhar, Panjkora, Bara, Chitral, Kalpani, Kurram, Kohat Toi, Gomal, and Tochi, which provide water for irrigation, power generation, and drinking. The province has a wide range of geographical factors that influence local climates, ranging from perennial freezing temperatures in the north to moderate winters and scorching summers in the south (Khan & Khan 2013). Climate change is popular in the province, as evidenced by unanticipated extreme weather events like riverine floods, strong winds, earthquakes, droughts, and so on. Zhang et al. (2008) projected the importance of the fluvial cycle in predicting climate change. Over the last three decades, climate change has produced an increase in the magnitude and frequency of floods (Zhang et al. 2011). The cumulative variability of fluvial overspills driven by climate change is a key source of worry in the literature (Dong et al. 2009). The Yangtze River in China, the Elbe in Germany, the Brahmaputra in Bangladesh, the Vistula and Oder in Poland, and the Indus in Pakistan, all experienced severe flooding during the first decade of the twenty-first century (2001–2010) (Chowdhury 2003; Gupta & Sah 2008; Wang et al. 2010; Lixin et al. 2012). Yu & Hu (2013) documented vegetation cover dynamics in distinct Chinese ecosystems, while Fabricante et al. (2009) identified climate conditions and land use changes as important contributing agents to vegetation dynamics.

Climate change is the world's most pressing concern at present. The climate has been warming since the beginning of the twentieth century, which is linked to excessive greenhouse gas (GHG) emissions (Crowley 2000) caused by fast industrialization and urbanization. Climate change estimates for South Asia in the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5) show warming above the world mean. According to climate model projections, the earth's land surface temperature is likely to rise 0.3–1.7 °C during the twenty-first century for the lowest GHG emissions with transient mitigation and 2.6–4.8 °C for business as usual carbon-intensive emissions. There is a need to reduce global carbon emissions by 40–70% by 2050 compared with 2010 levels, and to near-zero emissions by the end of the century. The current drive for power and wealth by the world's wealthy economies is compounding the crisis by releasing alarming volumes of greenhouse gases into the sky, putting developing countries like Pakistan in jeopardy. As the KP province of Pakistan is more exposed to dangers induced by climate change, the province requires more attention in research projects to alleviate the effects of the climate crisis. Nowshera district, located near the Indus and Kabul rivers, is prone to severe flooding and has been severely hit by heavy monsoon rainfall. District Nowshera as part of the Peshawar valley/basin demands more attention owing to the province's reliance on agricultural production because the valley is deemed the KP's food basket. The current research will (1) assess climate parameters variation trends for validation of the phenomenon of climate change in the study area, (2) track the spatiotemporal change in vegetation cover driven by climate change, (3) perform regression analysis of NDVI and environmental variables such as temperature, precipitation, relative humidity, and solar radiation, and (4) examine time-series NDWI data of the research region for hydrological fluctuations.

Data from the Shuttle Radar Topography Mission (SRTM) Digital Elevation Models (DEMs) were obtained from USGS Earth Explorer. The elevation data has a spatial resolution of 1 arc-second, which is approximately equal to 30 m.

It is a globally accepted fact that climate is changing as nature keep changing over time. To assess the trend of climate change in the study area, data obtained from regional center, Pakistan Meteorological Department, Peshawar (temperature, precipitation, and relative humidity), and NASA Power Data Access Viewer (solar radiation) was properly sorted, tabulated for monthly and seasonal arrangement in Excel and then analyzed for time series trends by applying modified Mann-Kendall trend and Sen's slope statistics using R software package version 4.2.0.

Climate change has a significant impact on NDVI. The change in NDVI caused by climatic conditions was investigated using simple and multivariate linear regression models. Annual average temperature (°C), precipitation (in), relative humidity (%), and solar radiation were the climate characteristics plotted against NDVI for their correlation. The analyses were carried out in R, and scatter plots were graphed for data visualization and comprehension.

For spatial variation trend analysis, the ArcGIS 10.8 software package was used. The averaged NDVI imagery was processed for 20 years of NDVI variation, which indicated a composite change in vegetation cover across the research period. Processing NDVI images from the rainy/wet (August) and dry (June) months of each year during the study period resulted in a vivid spatial variation trend depiction. Supervised maximum likelihood classification was used to build a training data layer file with high accuracy. The training layer classes were applied to corresponding NDVI imagery, which clearly portrayed and contrasted spatial variation in the vegetation cover of each year's wet and dry months.

Spatial analysis of satellite images for the months preceding floods and flood month was performed in the ArcGIS 10.8 package to exhibit the delineation and flooding of water bodies in the research area. For NDWI and MNDWI data, Landsat imagery from the study area's floods in 2005 and 2010 was used. The MNDWI was more accurate in water body visualizations, so it was processed further for the purpose. A training data layer file was created by supervised maximum likelihood classification of an established MNDWI imagery and was used to imageries from the months under consideration for visualization of floods and post-flood impacts on the study area.

Working on SRTM DEM data in ArcGIS, the research region was further probed for flood risk. Watershed delineation for the research region was created by processing the aforementioned dataset, which included elevation, water pathways, and flood inundation areas.

Climate change is colloquially known as global warming due to the profound and prominent rise in climate factor, i.e. temperature over time. The most recent global climate record reports period of 1979–97 as the warmest period due to about 0.6 °C increase in the earth's surface average temperature (Knutson et al. 2010). However, the change in rate and pattern of precipitation, relative humidity, and solar radiation are also reflecting climate change being climate parameters.

Table 1 shows the results of modified Mann-Kendall trend test and Sen's slope analysis for 20 years (2001–2020) monthly climate data performed in R. The average monthly temperature is although reflecting no trend at 95% confidence, but is actually increasing as evident from monthly maximum temperature statistics with significant results. The monthly average temperature is the average of monthly maximum and minimum temperatures in which the maximum temperature is showing positive (increasing) trend, while the monthly minimum temperature negative (decreasing) trend, the average thus causes no trend in the average temperature. Furthermore, the seasonal data analyzed in R shows highly significant increase in monthly average temperature (summer) besides monthly maximum temperature (summer).

The study area witnessed an irregular pattern of precipitation for the last decades. As a part of the monsoon region, the area receives heavy monsoon rains during the monsoon period but with significantly different rates and patterns each year. Table 1 depicts no trend for the monthly precipitation data of overall year for the study period due to highly irregular patterns and rates. The area faces severe droughts due to the changed pattern and rate of precipitation which has huge economic and societal impacts. While sometimes there are intensive rainfalls causing fluvial and flash floods. The seasonal precipitation data for winter reveals a significant decrease in the months of December, January, and February (winter). The decrease has caused losses in agriculture for wheat crops in rain-fed areas.

Table 1 reveals a significant increase in relative humidity during the study period which may have additive effect in warming the climate of the study area. Water vapors in air act as GHG, hence have a negative impact on the study area. However, it has a positive correlation with NDVI (vegetation) as Figure 5 depicts. Solar radiation shows a non-significant negative trend as Table 1 represents. The decreasing trend of the climate factor has a negative impact on vegetation (NDVI) (Figure 6) due to decline of photosynthetic potential of plants.

The analysis demonstrates a strong link between climatic parameters and NDVI. The NDVI has a positive correlation and regression coefficients when measured over time, therefore it improves with certain irregular fluctuations. Climate variables (precipitation and relative humidity) have a significant beneficial impact on NDVI. Temperature and solar radiation, on the other hand, have a negative correlation with NDVI. Figure 2 depicts the relationship between NDVI and time, temperature, precipitation, relative humidity, and solar radiation in such a way that all of these variables are represented both to be dependent (row-wise) and independent (column-wise) simultaneously. The multiple linear regression in R demonstrated the variables to be significantly correlated with NDVI while subsequently reveals the extent of data interrelationship and model prediction.

The following five stages can be seen in the variation trend from 2001 to 2020 (Figure 4). From 2001 to 2003, the first stage exhibits an ascending trend with a steep pattern. The years after that, from 2004 to 2007, show a minor variance with no discernible trend. Between 2007 and 2010, there was a declining trend in vegetation cover in the research area, which could be ascribed to harsh weather conditions owing to regional climate change. NDVI fluctuated the greatest in the fourth stage, from 2011 to 2016, showing highest annual changes and a major change in vegetation cover. This stage of NDVI temporal variation showed a net rising tendency, with years of significant falls in NDVI values alternating. The odd years have greater NDVI values, whereas the even years, such as 2012, 2014, and 2016, have lower NDVI values, indicating a trend in vegetation cover. The NDVI values dramatically increased in the latter stage of the research period, from 2017 to 2020, suggesting the quickest upward growing trend. The vegetation improved dramatically in the final stage due to the Pakistan government's ‘billion tree tsunami’ project in KP. The general trend of NDVI variation revealed a gradual increase in vegetation, implying a significant improvement in the studied area's vegetation cover during the last 20 years.

The climate in the district of Nowshera is generally arid and semi-arid. However, due to harsh weather conditions, i.e. natural catastrophes caused by climate change, the area's aridity status has altered significantly in the past. According to the NDVI values representing the area's vegetation cover, the aridity has significantly decreased over the last 20 years, hence improving vegetation. The area's overall vegetation cover has improved marginally since 2001, however two sections remain parched. The western and south-western areas of the research area, namely Jalozai, Haji Yar Khan Banda, Khaisari, Lakari, and Spin Kana Kalan (see Figure 1 for locating these areas) exhibit a persistent arid character (NDVI = 0.25) with a modest improvement each year (Figure 5). The area's low NDVI values are owing to the region's high priority for urban land, as the area is primarily bare ground with gravel sandy soil and low in vegetation and agricultural productivity. Because of the dense human population, rapid human development, and frequent human activities, the government is developing this region for housing projects and educational institutes. The study area's second main desert zone includes Tar Khel, Khai, Inzari, Khawarai, Mali Khel, and Mando Khel (see Figure 1) in the south-eastern corner. In addition, bare ground with sparse vegetation can be seen in the district's north-eastern region, near Misri Banda, Mughalki, Nandrak, Jalarona, and Raj Muhammad Kali. Akbarpura, Mohib Banda, Amankot, Zande Banda Chel, Haryana, Pir Sabaq, and Cherat (see Figure 1) are among the regions having NDVI values ranging from 0.4 to 0.65. These are either forested or agriculturally intensive areas (Figure 5). The research region is made up of 26.21% (NDVI = 0.25) bare land and urbanized land. The vegetation of farmlands, green pastures, rangelands, and shrublands of small shrubs, which make up a large portion of the study region (60.44%), is represented by NDVI values ranging from 0.25 to 0.45. The remaining 12.98% of the area with an NDVI > 0.45 contains dense vegetation of evergreen Prosopis juliflora and Lantana camara shrublands, some dense agricultural lands, and the Cherat Hills forest (Figure 5).

Figure 6 represents the NDVI of the greenest month of the year, August, which shows the district's highest vegetation average NDVI (range from 0.55 to 0.75) with a steady significant improvement year after year. The area's NDVI image in 2001 shows a significant amount of bare land, which steadily decreased over the years with some slight irregular oscillations, until the NDVI image in 2020 reveals a minuscule amount of bare land in the study area. The same month is inside the monsoon season, which means it sees a lot of rain and has a lot of humidity. As shown on the map, the month had the heaviest flash flood in the area's history, which occurred in 2010.

Several extreme weather events, such as riverine floods, earthquakes, storms, hail, and other climate-related events, occurred in the district during the study period; however, the floods were common. Using Landsat 5 and 7 MNDWI data, the current work delineates flash floods in 2005 and 2010. Due to the presence of the Kabul River at its northern territory and proximity and junction to the Indus River on its eastern border, the district is at high risk of riverine flooding. The floods have wreaked havoc on the district's infrastructure, resulting in massive economic losses as well as thousands of deaths. Fluvial floods from the Kabul River, on the other hand, deliver silt and sediment, making the upper layer of the soil fertile for some of the area's locations.

The fundamental determinant of an area's vegetation is its climate (Walter 2012). Climate change can be influenced by changes in vegetation cover (Zoran et al. 2016). Vegetation has been degraded as a result of human interference in the surrounding ecosystem (Yu et al. 2021). There is a clear link between climate and vegetation. Pakistan, with its diverse climate, is experiencing rapid climate change and weather variability. According to the Global Climate Risk Index 2021, it is among the top ten (bottom ten colloquially) countries most vulnerable to climate-related extreme weather (Eckstein et al. 2021). Climate change has had both positive and negative impacts on the research area, according to the current analysis. The phenomenon has a significant impact on the research area's vegetation cover. Natural processes and human activities cause climate parameters to vary in pattern and function (Dale 1997). Overall, the climate parameters, namely, precipitation, and relative humidity have a significant beneficial impact on vegetation cover (Zhou et al. 2019), whereas average temperature and solar radiation have a significant negative impact. In the last two decades (2001–2020), the vegetation cover has changed dramatically. Climate change and human action both have severely harmed the vegetation in a few specific locations. However, the vegetative status of the majority of the studied region has improved. The government's ‘billion tree tsunami’ campaign, Pakistan's greatest conservation project (Mumtaz & Ali 2019), is one of the key initiatives to counteract the detrimental effects of the climate change. Climate change is having a significant impact on the research area's weather extremes (Coumou & Rahmstorf 2012). Droughts, heatwaves, high winds, torrential rains, hailstones, and storms are the most common meteorological events in the study region, while flash and fluvial floods are the worst extreme event so far witnessed. Floods have harmed the economic and potential yield of the study region, as well as resulting in a large number of human deaths. They, on the other hand, transported alluvium and silt (Jacobsen & Adams 1958) changing the area's aridity status. The alluvium gave a fertile layer to parts of the study area's bare patches, resulting in a positive output. Furthermore, the floods transported seeds of non-indigenous species (Merriam 2003), which added to the district's flora. Parthenium hysterophorus L., Emex australis Steinh., Prosopis juliflora (Sw.) Dc., and Lantana camara L. are examples of non-native, exotic plant species that have infested the natural flora. These invasive plant species have successfully supplanted a number of native plant species as a result of their intense competition (Knight et al. 2009). The invasion of the formidable opponent, Parthenium hysterophorus L., reduces the density and spread of the native Cannabis sativa L. The study concludes that climate change has occurred in the studied area, with both positive and negative consequences. In a nutshell, climate change exacerbates weather extremes, resulting in massive economic and human losses (Ciscar et al. 2011) with significant impact on vegetation and land covers of the region. The study's findings will benefit in the decision-making process for vegetation restoration, regional sustainability, and climate change mitigation strategies.

Regional Meteorological Center, Pakistan Meteorological Department, Peshawar, Pakistan, provided meteorological data. This article is part of the Ph.D. thesis of the corresponding author.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.