ABSTRACT
Climate change's impact on wetlands has heightened concerns regarding the growing incidence of drought. It is imperative to continuously monitor long-term changes in wetlands to detect variations. This study evaluates the enduring variability of remote sensing indicators within 25 watershed areas in Algeria, recognized for their substantial biodiversity. We employed two robust statistical techniques (linear regression and the Mann-Kendall test) with data from diverse sources, including MODIS satellite data. From a time-series dataset spanning 22 years, we assembled several pivotal indicators: normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), normalized difference water index (NDWI), and land surface temperature (LST), selected to evaluate vegetation and water stress within the study areas. Our analysis revealed that NDVI showed more conspicuous temporal response when compared with EVI and NDWI. We uncovered negative correlations between NDVI and LST, underscoring the influence of drought and plant stress on vegetation within the study areas (R² = 0.109 to R² = 0.5701). Our Mann-Kendall trend analysis underscored the significance of NDVI and its robust association with EVI and NDWI. Understanding the dynamics of vegetation and water stress is of paramount importance for projecting changes in ecosystems, particularly within water bodies subject to climate variability.
HIGHLIGHTS
Study reveals increasing drought events in water bodies due to climate change.
Assessing remote sensing indicators in 25 biodiverse watersheds aids biodiversity conservation efforts.
The research underscores the crucial role of remote sensing technology in identifying fluctuations and aiding biodiversity conservation.
Understanding vegetation and water stress dynamics is vital for accurate ecosystem forecasts.
INTRODUCTION
The Earth's ecosystem has undergone rapid and extensive transformations over the past five decades, representing unprecedented changes not previously observed. Vegetation plays a foundational role within this ecosystem, and alterations in vegetation have far-reaching implications for global climate dynamics and the carbon cycle. A substantial body of evidence supports the assertion that nearly all terrestrial ecosystems are experiencing degradation, largely attributed to human activities and the pervasive impacts of global warming (Stellmes et al. 2010; Yang et al. 2017).
Remote sensing has significantly transformed our perception, utilization, and management of land assets (Huang et al. 2020). Numerous Earth observation programs have facilitated the generation of increasingly comprehensive insights into the wide array of natural variations and anthropogenic activities taking place on the Earth's surface. This is achieved through spectral analysis of the environment, enabling us to decipher valuable information from the data. The recording of data in time series has played a pivotal role in disseminating this knowledge (Potter et al. 2003).
Remotely sensed multi-spectral imagery contains bands to create a composite image that can be viewed and analyzed (Xue & Su 2017). Advancements in robust and widely accessible remote sensing imagery enable the development of new methodologies for evaluating assessments at broad, medium, and fine scales. This allows for the detection of persistent structural and functional environmental transformations (Nagendra et al. 2013; O'Connor et al. 2015). A variety of tools and methodologies have been developed and rigorously assessed to gauge the condition of natural resources, encompassing their responses to climate change (CC), extreme events, and human-induced stresses. Among these methods, geospatial techniques, notably remote sensing (RS), have been instrumental. Groundwater sustainability tools (GSTs) have been extensively employed in diverse research initiatives, providing insights into various aspects of water resources. For instance, Moumane et al. (2021) utilized GSTs to assess the state of groundwater resources. This approach enables a comprehensive evaluation of the health and sustainability of groundwater reservoirs. In other studies, the integration of geographic information system (GIS) and RS techniques, as demonstrated by Ben Salem et al. (2019) and Ougougdal et al. (2020), facilitated the examination of water erosion.
From a scientific perspective, there exists a diverse array of methodologies available for investigating the repercussions of CC on drought events. Among these, the normalized difference vegetation index (NDVI) stands as a commonly utilized tool for assessing the impacts of both drought and CC. One notable study conducted by Nanzad et al. (2019) effectively harnessed the NDVI anomaly as a tool for assessing drought conditions and investigating their relationship with various climatic factors over the period spanning from 2000 to 2016. In a related endeavor, Xie & Fan (2021) developed drought indicators by leveraging vegetation indices, specifically the NDVI and enhanced vegetation index (EVI). These indices were derived from data acquired through the Moderate Resolution Imaging Spectroradiometer (MODIS), offering a comprehensive view of vegetation changes in response to drought. Furthermore, the normalized difference water index (NDWI) has emerged as a significant tool for the analysis of hydrological droughts, as elucidated by Noorisameleh et al. (2020). This index aids in assessing variations in water content and availability, which are crucial aspects of hydrological drought events.
The indices, including the NDVI, NDWI, and EVI, are widely employed for surface observations. Their utility in evaluating the impacts of CC on drought conditions underscores their importance in advancing our understanding of these complex environmental phenomena.
The NDVI is a credible indicator of broad-scale transformations in vegetation density (Myneni et al. 1995; Wang et al. 2018). It has been extensively used for monitoring the vegetation dynamics in surveillance and estimating the vegetation feedback on a global scale (Nemani et al. 2003; Zhu et al. 2016) and regional scales (Piao et al. 2003; Wu et al. 2015). Over the last few decades, the global climate has witnessed some dramatic shifts. Furthermore, global warming has been shown to be undeniable (Karl & Trenberth 2003; IPCC 2023). Accordingly, in recent decades, there has been an increasing emphasis on the significant modifications observed in the Earth's climate and their implications for greening development (De Jong et al. 2011; Kong et al. 2020; Guo et al. 2021).
In the last several years, Algeria has suffered from the deterioration of its wetlands (Nedkov et al. 2018), confirming the results of previous studies (Sitayeb & Benabdeli 2008; Belgherbi & Benabdeli 2010; Ghodbani & Amokrane 2013; Megharbi et al. 2016; Souidi et al. 2016; Keria et al. 2023). Algeria comprises 1,451 wetlands, consisting of 762 natural areas and 689 artificial areas (some of which are included in the list of globally significant wetlands) (DGF 2017). In 2009, 42 Ramsar sites were classified, from the international agreement adopted in 1971 that deals with wetlands of international importance, as essential habitats for the Convention on Biological Diversity. This treaty recognizes the ecological, economic, and cultural importance of wetlands, thereby encouraging the preservation and sustainable use of these diverse ecosystems (Ramsar 2015), spanning a combined area of 2.960 million hectares. Among these conserved sites, saline lakes account for 45.23% of the coverage, estimated at roughly 2.07890 million hectares (Koopmanschap et al. 2011). These wetlands are typically found in coastal areas or valleys conducive to socio-economic growth, making them susceptible to being targeted for urban, agricultural, or industrial development. Their worldwide surface area has diminished by 64% and 71%, respectively, since 1900 (Davidson 2014). It should be noted that the weakness or lack of a management system for watersheds in Algeria significantly affects decisions and professionals involved in the management of wetland environments.
As the existing body of study in this particular field demonstrates a dearth of current investigations that employ rigorous statistical (remote sensing) approaches in a susceptible region affected by CC and anthropogenic activities. In light of this deficiency, in this study, the main focus is on understanding the spatial and temporal changes in 25 wetland areas located in Algeria, given their environmental and economic significance. Among the specific objectives of the study are the following: (1) to examine the temporal pattern of EVI, NDVI, NDWI, and LST indicators in a variety of water catchments in Algeria during the period from 2000 to 2022; (2) assessment of the influence of CCs on Algeria's watersheds using Google Earth Engine (GEE) data; (3) identification of environmentally sensitive areas in Algeria's watersheds.
MATERIALS AND METHODS
Study area
Characteristics of all the study areas
Watershed . | Watershed area (ha) . | Water quality . | Hydrological cadency . | Current status . | Latitude (N) . | Longitude (E), (W) . | Climate . |
---|---|---|---|---|---|---|---|
Macta Marsh | 44,500 | Salty | Permanent | Ramsar (2001) | 35.704670° | -0.121092°W | Semi-arid |
Great Sebkha | 56,870 | Salty | Permanent | Ramsar (2001) | 35.518985° | -0.860435°W | Semi-arid |
Telamine Lake | 2,399 | Salty | Permanent | Ramsar (2004) | 35.736568° | -0.382841°W | Semi-arid |
Arzew saline | 5,778 | Salty | Permanent | Ramsar (2004) | 35.860485° | -0.329562°W | Semi-arid |
Boughezoul dam | 9,058 | Salty | Permanent | Ramsar (2011) | 35.738283° | 2.785820°E | Semi-arid |
Chott Zehrez Chergui | 50,985 | Salty | Permanent | Ramsar (2003) | 34.944018° | 2.862758°E | Semi-arid |
Chott Ech Chergui | 855,500 | Salty | Permanent | Ramsar (2001) | 34.350000° | 0.500000°E | Semi-arid |
Dayet El Ferd | 3,323 | Salty | Permanent | Ramsar (2004) | 34.497783° | -1.245122°W | Semi-arid |
Chott Zehrez Gharbi | 52,200 | Salty | Permanent | Ramsar (2003) | 34.944018° | 2.862758°E | Semi-arid |
Chott El Hodna | 362,000 | Salty | Permanent | Ramsar (2001) | 35.442812° | 4.797680°E | Semi-arid |
Chott Melghir | 551,500 | Salty | Irregular | Ramsar (2002) | 34.122969° | 6.001930°E | Arid |
Beni Bahdel dam | 54,630 | Salty | Irregular | Not protected | 34.703221° | -1.500366°W | Semi-arid |
Bougara dam | 11,320 | Salty | Irregular | Not protected | 35.565565° | 1.908756°E | Semi-arid |
Boughrara dam | 175,450 | Salty | Irregular | Not protected | 34.874073° | -1.653459°W | Semi-arid |
Bouhanifia dam | 34,520 | Salty | Irregular | Not protected | 35.278106 ° | -0.071455°W | Semi-arid |
Chorfa dam | 70,210 | Salty | Irregular | Not protected | 35.384663° | -0.274070°W | Semi-arid |
Cheliff dam | 50,000 | Salty | Irregular | Not protected | 35.990491° | 0.446296°E | Semi-arid |
Chott Merouane | 337,700 | Salty | Irregular | IBA-Ramsar (2001) | 33.5328° | 6.1034°E | Arid |
Dahmoni dam | 39,520 | Salty | Irregular | Not protected | 35.451481° | 1.449635°E | Semi-arid |
Gargar dam | 358,280 | Salty | Irregular | Not protected | 35.913654° | 1.007704°E | Semi-arid |
Karrada dam | 65,000 | Salty | Irregular | Not protected | 36.058190° | 0.384113°E | Semi-arid |
Ksob dam | 22,720 | Salty | Irregular | Not protected | 35.838615° | 4.565161°E | Arid |
Sarno dam | 21,250 | Salty | Irregular | Not protected | 35.294987° | -0.594364°W | Semi-arid |
El Golea | 18,947 | Salty | Permanent | Ramsar (2004) | 30.625869° | 2.875530°E | Arid |
Chott Ain El Beida | 6,853 | Salty | Irregular | Ramsar (2004) | 31.800687° | 5.494661°E | Arid |
Watershed . | Watershed area (ha) . | Water quality . | Hydrological cadency . | Current status . | Latitude (N) . | Longitude (E), (W) . | Climate . |
---|---|---|---|---|---|---|---|
Macta Marsh | 44,500 | Salty | Permanent | Ramsar (2001) | 35.704670° | -0.121092°W | Semi-arid |
Great Sebkha | 56,870 | Salty | Permanent | Ramsar (2001) | 35.518985° | -0.860435°W | Semi-arid |
Telamine Lake | 2,399 | Salty | Permanent | Ramsar (2004) | 35.736568° | -0.382841°W | Semi-arid |
Arzew saline | 5,778 | Salty | Permanent | Ramsar (2004) | 35.860485° | -0.329562°W | Semi-arid |
Boughezoul dam | 9,058 | Salty | Permanent | Ramsar (2011) | 35.738283° | 2.785820°E | Semi-arid |
Chott Zehrez Chergui | 50,985 | Salty | Permanent | Ramsar (2003) | 34.944018° | 2.862758°E | Semi-arid |
Chott Ech Chergui | 855,500 | Salty | Permanent | Ramsar (2001) | 34.350000° | 0.500000°E | Semi-arid |
Dayet El Ferd | 3,323 | Salty | Permanent | Ramsar (2004) | 34.497783° | -1.245122°W | Semi-arid |
Chott Zehrez Gharbi | 52,200 | Salty | Permanent | Ramsar (2003) | 34.944018° | 2.862758°E | Semi-arid |
Chott El Hodna | 362,000 | Salty | Permanent | Ramsar (2001) | 35.442812° | 4.797680°E | Semi-arid |
Chott Melghir | 551,500 | Salty | Irregular | Ramsar (2002) | 34.122969° | 6.001930°E | Arid |
Beni Bahdel dam | 54,630 | Salty | Irregular | Not protected | 34.703221° | -1.500366°W | Semi-arid |
Bougara dam | 11,320 | Salty | Irregular | Not protected | 35.565565° | 1.908756°E | Semi-arid |
Boughrara dam | 175,450 | Salty | Irregular | Not protected | 34.874073° | -1.653459°W | Semi-arid |
Bouhanifia dam | 34,520 | Salty | Irregular | Not protected | 35.278106 ° | -0.071455°W | Semi-arid |
Chorfa dam | 70,210 | Salty | Irregular | Not protected | 35.384663° | -0.274070°W | Semi-arid |
Cheliff dam | 50,000 | Salty | Irregular | Not protected | 35.990491° | 0.446296°E | Semi-arid |
Chott Merouane | 337,700 | Salty | Irregular | IBA-Ramsar (2001) | 33.5328° | 6.1034°E | Arid |
Dahmoni dam | 39,520 | Salty | Irregular | Not protected | 35.451481° | 1.449635°E | Semi-arid |
Gargar dam | 358,280 | Salty | Irregular | Not protected | 35.913654° | 1.007704°E | Semi-arid |
Karrada dam | 65,000 | Salty | Irregular | Not protected | 36.058190° | 0.384113°E | Semi-arid |
Ksob dam | 22,720 | Salty | Irregular | Not protected | 35.838615° | 4.565161°E | Arid |
Sarno dam | 21,250 | Salty | Irregular | Not protected | 35.294987° | -0.594364°W | Semi-arid |
El Golea | 18,947 | Salty | Permanent | Ramsar (2004) | 30.625869° | 2.875530°E | Arid |
Chott Ain El Beida | 6,853 | Salty | Irregular | Ramsar (2004) | 31.800687° | 5.494661°E | Arid |
Data and processing
The data analyzed in this research were collected from the GEE platform (https://developers.google.com/earth-engine/datasets/). The platform hosts a vast array of remote sensing datasets and is equipped with robust data processing capabilities (Gorelick et al. 2017). In this study, we utilized a dataset that includes MODIS. The research period for this work spans from 2000 to 2022.
The MOD11A1, MOD09GA, and MOD13Q1 products were utilized in our study, as these datasets provide essential information for variables such as the NDVI, EVI, NDWI, and land surface temperature (LST). The sources and characteristics of the various datasets utilized in this work are listed in Table 2.
Datasets utilized in this study
Data products . | Data types . | Spatial resolution . | Temporal resolution . | Spatial coverage . | Duration . |
---|---|---|---|---|---|
MOD09GA | NDWI | 250 m | 16 days | Global | 2000–2022 |
MOD13Q1 | NDVI/EVI | 250 m | 16 days | Global | 2000–2022 |
MOD11A1 | LST | 1,000 m | 8 days | Global | 2000–2022 |
Data products . | Data types . | Spatial resolution . | Temporal resolution . | Spatial coverage . | Duration . |
---|---|---|---|---|---|
MOD09GA | NDWI | 250 m | 16 days | Global | 2000–2022 |
MOD13Q1 | NDVI/EVI | 250 m | 16 days | Global | 2000–2022 |
MOD11A1 | LST | 1,000 m | 8 days | Global | 2000–2022 |
Method
Assessment of NDVI temporal dynamics
The NDVI is utilized to quantify the variability of visible and near-infrared light in vegetation. This is achieved by calculating the difference between these spectral bands within specific wavelength ranges and normalizing it across the entire spectrum. The NDVI is calculated using the equation (NIR−VIS)/(NIR + VIS), where NIR represents the spectral reflectance in the near-infrared band and VIS represents the reflectance in the visible band. These indicators have been extensively employed to capture the dynamics and growth of vegetation (Tucker et al. 1986; Myneni et al. 1997; Nemani et al. 2003; Piao et al. 2011; Donohue et al. 2016).
In the provided equation, ‘Slope’ represents the change in NDVI over time;
is the NDVI value calculated using the formula for the i-th year; and n represents the total number of years during the assessment duration. The ‘Slope’ indicates the rate of change in vegetation growth. If the Slope > 0, it suggests an increase in vegetation development with a decrease in vegetation coverage. When the Slope = 0, it indicates stable and consistent regional vegetation growth. On the other hand, if the Slope < 0, it suggests a decrease in vegetation development activity with an increase in vegetation coverage (Zhe & Zhang 2021).
Determination of different remote sensing indices
Here, and
are low- and high-resolution spectral indexes, respectively;
and
indicate LST at low and high resolutions, respectively (Xu et al. 2021).
Mann–Kendall trend analysis

RESULTS
The evolving pattern of NDVI in the 25 areas over the course of the last 22 years
The pattern of interannual changes in the annual average NDVI for 25 watersheds in Algeria over the period of 2000 to 2022.
The pattern of interannual changes in the annual average NDVI for 25 watersheds in Algeria over the period of 2000 to 2022.
Specifically, elevated NDVI values, ranging between 0.4 and 0.6, were observed in certain areas, while lower values, predominantly ranging from 0 to 0.2, were more prevalent. The Sarno watershed stood out with the highest NDVI value in 2001, exceeding 0.6, while the Chott Merouane watershed recorded the lowest NDVI value of 0.07 in 2006.
In the upper reaches of the region, the 22-year average NDVI was 0.4, indicating a moderate overall vegetation level. There was an upward trend in vegetation from inland areas towards the northwest. Among specific areas, the Chott Merouane area in the northeast of the North Sahara had the lowest vegetation proportion, followed by the Golea Basin, Chott Melrhir, Ain Beida, Chott El Hodna, and Ksob dam. In contrast, the highest vegetation ratios were observed in areas such as the Sarno dam, Telamine Lake, Arzew Saline, Macta Marshes, Great Sebkha, Boughrara dam, Cheliff dam, Chorfa dam, Beni Bahdel dam, Karrada dam, and Bouhanifia dam, located in the northwest. Bougara dam, Dahmoni dam, and Gargar dam, situated in the central and upper reaches of the region, also had relatively high vegetation cover, with an annual average NDVI for these areas reaching 0.6, indicating medium to high vegetation levels.
Conversely, areas from Ksob dam off the southern saline of Golea had the lowest average NDVI, measuring 0.05. These areas were predominantly classified as having moderate to low or low vegetation cover, suggesting sparse vegetation and unfavorable environmental conditions. Hence, there is a need for implementing appropriate measures to increase green spaces in these regions. The average NDVI for Boughezoul dam, Chott Zehrez Chergui, Chott Ech Chergui, Chott Zehrez Gharbi, and Dayet El Ferd ranged from 0.2 to 0.4, indicating relatively low vegetation cover. These basins can be categorized as medium/low vegetation areas with a very poor vegetation condition.
Spatio-temporal variation of NDVI, EVI, NDWI, and LST
Interannual fluctuations in annual average NDWI across 25 Algerian watersheds (2000–2022).
Interannual fluctuations in annual average NDWI across 25 Algerian watersheds (2000–2022).
Interannual fluctuations in annual average EVI across 25 Algerian watersheds (2000–2022).
Interannual fluctuations in annual average EVI across 25 Algerian watersheds (2000–2022).
Interannual fluctuations in annual average LST across 25 Algerian watersheds (2000–2022).
Interannual fluctuations in annual average LST across 25 Algerian watersheds (2000–2022).
The associations between NDVI and LST
Annual assessment of relationship between NDVI and LST indices during 2000–2022.
Annual assessment of relationship between NDVI and LST indices during 2000–2022.
In other regions, a strong negative relationship between NDVI and LST has decreased over time. The R2 values in these areas range from 0.109 to 0.3793. This indicates that as LST values increase, there is a corresponding decrease in vegetation. The negative relationship between NDVI and LST suggests that higher temperatures are linked to reduced vegetation activity in these regions. It is essential to note that the connection between NDVI and LST differs for water bodies. In the case of water bodies or wetlands, a negative NDVI indicates the presence of water. The wave-like pattern observed between NDVI and LST for water bodies may be influenced by variables such as water level and temperature.
Overall, the observed relationships between NDVI and LST provide valuable insights into the dynamics of vegetation and LST in the study site. These findings are crucial for understanding the impact of temperature on vegetation health and for monitoring environmental changes over time.
Mann–Kendall analysis trends
The MK analysis was performed on the NDVI, EVI, NDWI, and LST datasets at a 95% confidence level, with serial correlation applied. The results of the MK test for the NDVI data are summarized in Table 3. Among the 25 regions examined, 15 displayed negative MK statistics, indicating an overall decreasing trend in NDVI over the study period. The associated p-values for these regions exceeded the specified significance level (alpha), indicating significant evidence of a monotonic decline in vegetation activity within these areas.
Findings of the two-tailed test for seasonality in the NDVI time series using the MK method
Study areas . | tau . | p-value . | Sen's slope . |
---|---|---|---|
Ain Beida | 0.167 | 0.00013 | 0.0004 |
Arzew saline | 0.016 | 0.700 | 0.006 |
Boughezoul dam | −0.007 | 0.871 | 0.005 |
Chott Zehrez Chergui | 0.014 | 0.732 | 0.000 |
Chott Ech Chergui | 0.136 | 0.001 | 0.006 |
Dayet El Ferd | −0.029 | 0.480 | −0.001 |
Chott Zehrez Gharbi | 0.021 | 0.621 | 0.002 |
Chott El Hodna | 0.009 | 0.829 | −0.001 |
Macta Marsh | −0.028 | 0.490 | 0.004 |
Chott Melrhir | 0.088 | 0.036 | 0.000 |
Great Sebkha | 0.049 | 0.234 | 0.004 |
Telamine Lake | 0.053 | 0.203 | 0.004 |
Beni Bahdel dam | 0.127 | 0.002 | 0.017 |
Bougara dam | 0.070 | 0.091 | 0.007 |
Boughrara dam | 0.049 | 0.239 | 0.014 |
Bouhanifia dam | 0.052 | 0.208 | 0.014 |
Chorfa dam | 0.024 | 0.560 | 0.009 |
Cheliff dam | 0.138 | 0.001 | 0.012 |
Chott Merouane | −0.047 | 0.262 | 0.000 |
Dahmoni dam | 0.092 | 0.026 | 0.007 |
Golea | 0.344 | <0.0001 | 0.001 |
Gargar dam | 0.114 | 0.006 | 0.012 |
Karrada dam | 0.089 | 0.031 | 0.002 |
Ksob dam | 0.162 | <0.0001 | 0.005 |
Sarno dam | −0.002 | 0.960 | −0.001 |
Study areas . | tau . | p-value . | Sen's slope . |
---|---|---|---|
Ain Beida | 0.167 | 0.00013 | 0.0004 |
Arzew saline | 0.016 | 0.700 | 0.006 |
Boughezoul dam | −0.007 | 0.871 | 0.005 |
Chott Zehrez Chergui | 0.014 | 0.732 | 0.000 |
Chott Ech Chergui | 0.136 | 0.001 | 0.006 |
Dayet El Ferd | −0.029 | 0.480 | −0.001 |
Chott Zehrez Gharbi | 0.021 | 0.621 | 0.002 |
Chott El Hodna | 0.009 | 0.829 | −0.001 |
Macta Marsh | −0.028 | 0.490 | 0.004 |
Chott Melrhir | 0.088 | 0.036 | 0.000 |
Great Sebkha | 0.049 | 0.234 | 0.004 |
Telamine Lake | 0.053 | 0.203 | 0.004 |
Beni Bahdel dam | 0.127 | 0.002 | 0.017 |
Bougara dam | 0.070 | 0.091 | 0.007 |
Boughrara dam | 0.049 | 0.239 | 0.014 |
Bouhanifia dam | 0.052 | 0.208 | 0.014 |
Chorfa dam | 0.024 | 0.560 | 0.009 |
Cheliff dam | 0.138 | 0.001 | 0.012 |
Chott Merouane | −0.047 | 0.262 | 0.000 |
Dahmoni dam | 0.092 | 0.026 | 0.007 |
Golea | 0.344 | <0.0001 | 0.001 |
Gargar dam | 0.114 | 0.006 | 0.012 |
Karrada dam | 0.089 | 0.031 | 0.002 |
Ksob dam | 0.162 | <0.0001 | 0.005 |
Sarno dam | −0.002 | 0.960 | −0.001 |
The MK test outcomes for EVI, outlined in Table 4, mirrored the NDVI results in the majority of regions, suggesting a similar trend. However, an exception was noted in the Ksob region, where the EVI trend displayed a distinct positive direction, signifying an overall increase in vegetation activity. The p-value associated with this region was lower than the specified significance level (alpha), providing strong evidence of a significant monotonic upward trend in vegetation activity in the Ksob region (Figure 7).
Findings of the two-tailed test for seasonality in the EVI time series using the MK method
Study areas . | tau . | p-value . | Sen's slope . |
---|---|---|---|
Ain Beida | 0.142 | 0.001 | 0.0002 |
Arzew saline | −0.003 | 0.951 | 0.002 |
Boughezoul dam | −0.003 | 0.942 | 0.002 |
Chott Zehrez Chergui | 0.080 | 0.055 | 0.002 |
Chott Ech Chergui | 0.199 | <0.0001 | 0.002 |
Dayet El Ferd | 0.000 | 0.993 | −0.001 |
Chott Zehrez Gharbi | 0.051 | 0.220 | 0.001 |
Chott El Hodna | 0.068 | 0.103 | 0.001 |
Macta Marsh | −0.046 | 0.262 | 0.000 |
Chott Melrhir | 0.121 | 0.004 | 0.001 |
Great Sebkha | 0.025 | 0.548 | −0.002 |
Telamine Lake | 0.024 | 0.558 | −0.002 |
Beni Bahdel dam | 0.109 | 0.008 | 0.005 |
Bougara dam | 0.079 | 0.058 | 0.002 |
Boughrara dam | 0.036 | 0.387 | 0.007 |
Bouhanifia dam | 0.059 | 0.153 | 0.004 |
Chorfa dam | 0.003 | 0.947 | 0.000 |
Cheliff dam | 0.107 | 0.010 | 0.000 |
Chott Merouane | 0.069 | 0.105 | 0.006 |
Dahmoni dam | 0.089 | 0.031 | 0.001 |
Golea | 0.301 | <0.0001 | 0.005 |
Gargar dam | 0.109 | 0.008 | 0.002 |
Karrada dam | 0.086 | 0.036 | 0.005 |
Ksob dam | 0.010 | 0.803 | 0.005 |
Sarno dam | 0.010 | 0.803 | 0.000 |
Study areas . | tau . | p-value . | Sen's slope . |
---|---|---|---|
Ain Beida | 0.142 | 0.001 | 0.0002 |
Arzew saline | −0.003 | 0.951 | 0.002 |
Boughezoul dam | −0.003 | 0.942 | 0.002 |
Chott Zehrez Chergui | 0.080 | 0.055 | 0.002 |
Chott Ech Chergui | 0.199 | <0.0001 | 0.002 |
Dayet El Ferd | 0.000 | 0.993 | −0.001 |
Chott Zehrez Gharbi | 0.051 | 0.220 | 0.001 |
Chott El Hodna | 0.068 | 0.103 | 0.001 |
Macta Marsh | −0.046 | 0.262 | 0.000 |
Chott Melrhir | 0.121 | 0.004 | 0.001 |
Great Sebkha | 0.025 | 0.548 | −0.002 |
Telamine Lake | 0.024 | 0.558 | −0.002 |
Beni Bahdel dam | 0.109 | 0.008 | 0.005 |
Bougara dam | 0.079 | 0.058 | 0.002 |
Boughrara dam | 0.036 | 0.387 | 0.007 |
Bouhanifia dam | 0.059 | 0.153 | 0.004 |
Chorfa dam | 0.003 | 0.947 | 0.000 |
Cheliff dam | 0.107 | 0.010 | 0.000 |
Chott Merouane | 0.069 | 0.105 | 0.006 |
Dahmoni dam | 0.089 | 0.031 | 0.001 |
Golea | 0.301 | <0.0001 | 0.005 |
Gargar dam | 0.109 | 0.008 | 0.002 |
Karrada dam | 0.086 | 0.036 | 0.005 |
Ksob dam | 0.010 | 0.803 | 0.005 |
Sarno dam | 0.010 | 0.803 | 0.000 |
In NDWI, positive values observed in 18 regions suggest an increasing trend in water content. The corresponding p-values (>0.05) for these regions indicate a monotonically positive trend. Conversely, other regions with p-values below 0.05 show a significant negative trend in water content. MK test results provide evidence of monotonic trends, either increasing or decreasing, in NDWI values across the study regions (Figure 7).
The application of long-term sequence analysis (LST) with the MK test provides results from the statistical trend analysis of annual temperature degrees, as outlined in Table 6. These findings reveal a statistically significant positive trend in all wetlands, indicating a gradual increase over time. The analysis suggests a substantial alteration in annual temperature magnitudes, particularly between the years 2020 and 2022. The confirmed results, based on the MK statistics and accompanying p-values at significance levels of 5% and 1%, demonstrate statistical significance (Figure 7). As anticipated, both variables representing the maximum temperature show a consistent upward trend. In summary, the findings of the trend test demonstrate a statistically significant pattern of temperature rise in all regions.
Sen's slope was utilized to analyze time-series data of natural indicators from 2020 to 2022 (Tables 3–6). In all scenarios, the indicators exhibited a noticeable and continuous upward trend. There was an increase in vegetation density based on NDVI and EVI values in Beni Bahdel dam, Bouhanifia dam, Boughrara dam, and Gargar dam. Additionally, NDWI analysis indicated an increase in water content in Ksob dam, Boughezoul dam, and Chott Zehrez Gharbi. The rise in surface temperature over time suggests environmental conditions characterized by excessive heat, which may negatively impact vegetation density.
Findings of the two-tailed test for seasonality in the NDWI time series using the MK method
Study areas . | Tau . | p-Value . | Sen slope . |
---|---|---|---|
Ain Beida | 0.028 | 0.0521 | −0.0001 |
Arzew Saline | 0.050 | 0.227 | −0.002 |
Boughezoul dam | 0.080 | 0.055 | 0.001 |
Chott Zehrez Chergui | 0.069 | 0.098 | 0.0001 |
Chott Ech Chergui | −0.044 | 0.292 | −0.001 |
Dayet El Ferd | 0.036 | 0.384 | 0.0002 |
Chott Zehrez Gharbi | 0.048 | 0.251 | 0.001 |
Chott El Hodna | −0.026 | 0.527 | 0.000 |
Macta Swamp | 0.083 | 0.044 | 0.003 |
Chott Melrhir | 0.019 | 0.657 | 0.0003 |
Great Sebkha | 0.075 | 0.069 | 0.006 |
Telamine Lake | 0.106 | 0.010 | −0.0004 |
Beni Bahdel dam | 0.086 | 0.038 | 0.005 |
Bougara dam | 0.078 | 0.062 | 0.001 |
Boughrara dam | 0.035 | 0.401 | 0.001 |
Bouhanifia dam | 0.034 | 0.412 | 0.001 |
Chorfa dam | 0.015 | 0.712 | 0.002 |
Cheliff dam | 0.134 | 0.001 | 0.001 |
Chott Merouane | 0.014 | 0.738 | −0.001 |
Dahmoni dam | 0.088 | 0.034 | 0.006 |
Golea | −0.132 | 0.002 | −0.001 |
Gargar dam | 0.079 | 0.058 | −0.0006 |
Karrada dam | 0.057 | 0.171 | −0.0001 |
Ksob dam | 0.153 | 0.0001 | 0.002 |
Sarno dam | −0.020 | 0.621 | −0.001 |
Study areas . | Tau . | p-Value . | Sen slope . |
---|---|---|---|
Ain Beida | 0.028 | 0.0521 | −0.0001 |
Arzew Saline | 0.050 | 0.227 | −0.002 |
Boughezoul dam | 0.080 | 0.055 | 0.001 |
Chott Zehrez Chergui | 0.069 | 0.098 | 0.0001 |
Chott Ech Chergui | −0.044 | 0.292 | −0.001 |
Dayet El Ferd | 0.036 | 0.384 | 0.0002 |
Chott Zehrez Gharbi | 0.048 | 0.251 | 0.001 |
Chott El Hodna | −0.026 | 0.527 | 0.000 |
Macta Swamp | 0.083 | 0.044 | 0.003 |
Chott Melrhir | 0.019 | 0.657 | 0.0003 |
Great Sebkha | 0.075 | 0.069 | 0.006 |
Telamine Lake | 0.106 | 0.010 | −0.0004 |
Beni Bahdel dam | 0.086 | 0.038 | 0.005 |
Bougara dam | 0.078 | 0.062 | 0.001 |
Boughrara dam | 0.035 | 0.401 | 0.001 |
Bouhanifia dam | 0.034 | 0.412 | 0.001 |
Chorfa dam | 0.015 | 0.712 | 0.002 |
Cheliff dam | 0.134 | 0.001 | 0.001 |
Chott Merouane | 0.014 | 0.738 | −0.001 |
Dahmoni dam | 0.088 | 0.034 | 0.006 |
Golea | −0.132 | 0.002 | −0.001 |
Gargar dam | 0.079 | 0.058 | −0.0006 |
Karrada dam | 0.057 | 0.171 | −0.0001 |
Ksob dam | 0.153 | 0.0001 | 0.002 |
Sarno dam | −0.020 | 0.621 | −0.001 |
Findings of the two-tailed test for seasonality in the LST time series using the MK method
Study areas . | tau . | p-value . | Sen's slope . |
---|---|---|---|
Ain Beida | 0.056 | 0.120 | 0.023 |
Arzew saline | 0.068 | 0.074 | 0.010 |
Boughezoul dam | 0.068 | 0.077 | 0.040 |
Chott Zehrez Chergui | 0.068 | 0.077 | 0.063 |
Chott Ech Chergui | 0.072 | 0.067 | 0.044 |
Dayet El Ferd | 0.072 | 0.066 | 0.053 |
Chott Zehrez Gharbi | 0.056 | 0.117 | 0.045 |
Chott El Hodna | 0.056 | 0.117 | 0.017 |
Macta Swamp | 0.036 | 0.221 | 0.028 |
Chott Melrhir | 0.064 | 0.083 | 0.013 |
Great Sebkha | 0.060 | 0.110 | 0.003 |
Telamine Lake | 0.060 | 0.110 | 0.003 |
Beni Bahdel dam | 0.072 | 0.067 | 0.002 |
Bougara dam | 0.072 | 0.060 | 0.006 |
Boughrara dam | 0.084 | 0.044 | 0.005 |
Bouhanifia dam | 0.072 | 0.061 | 0.009 |
Chorfa dam | 0.072 | 0.067 | 0.014 |
Cheliff dam | 0.024 | 0.309 | 0.007 |
Chott Merouane | 0.032 | 0.249 | 0.038 |
Dahmoni dam | 0.056 | 0.109 | 0.000 |
Golea | 0.048 | 0.151 | 0.021 |
Gargar dam | 0.048 | 0.151 | 0.021 |
Karrada dam | 0.040 | 0.189 | 0.006 |
Ksob dam | 0.044 | 0.183 | 0.022 |
Sarno dam | 0.076 | 0.062 | 0.040 |
Study areas . | tau . | p-value . | Sen's slope . |
---|---|---|---|
Ain Beida | 0.056 | 0.120 | 0.023 |
Arzew saline | 0.068 | 0.074 | 0.010 |
Boughezoul dam | 0.068 | 0.077 | 0.040 |
Chott Zehrez Chergui | 0.068 | 0.077 | 0.063 |
Chott Ech Chergui | 0.072 | 0.067 | 0.044 |
Dayet El Ferd | 0.072 | 0.066 | 0.053 |
Chott Zehrez Gharbi | 0.056 | 0.117 | 0.045 |
Chott El Hodna | 0.056 | 0.117 | 0.017 |
Macta Swamp | 0.036 | 0.221 | 0.028 |
Chott Melrhir | 0.064 | 0.083 | 0.013 |
Great Sebkha | 0.060 | 0.110 | 0.003 |
Telamine Lake | 0.060 | 0.110 | 0.003 |
Beni Bahdel dam | 0.072 | 0.067 | 0.002 |
Bougara dam | 0.072 | 0.060 | 0.006 |
Boughrara dam | 0.084 | 0.044 | 0.005 |
Bouhanifia dam | 0.072 | 0.061 | 0.009 |
Chorfa dam | 0.072 | 0.067 | 0.014 |
Cheliff dam | 0.024 | 0.309 | 0.007 |
Chott Merouane | 0.032 | 0.249 | 0.038 |
Dahmoni dam | 0.056 | 0.109 | 0.000 |
Golea | 0.048 | 0.151 | 0.021 |
Gargar dam | 0.048 | 0.151 | 0.021 |
Karrada dam | 0.040 | 0.189 | 0.006 |
Ksob dam | 0.044 | 0.183 | 0.022 |
Sarno dam | 0.076 | 0.062 | 0.040 |
The results of the Sen's slope analysis revealed that the most affected areas in terms of vegetation cover are Dayet El Ferd, Sarno dam, Chott El Hodna, Ain Beida, Arzew Saline, Chott Ech Chergui, Telamine Lake, Chott Merouane, Golea, Gargar dam, Karrada dam, and Sarno dam. The trend test results, with a significance level of 1%, showed statistically significant trends for the indicators, with the results remaining statistically significant at a 95% confidence level from 2020 to 2022.
Discussion
The primary aim of this study was to assess fluctuations in the NDVI and variations in water bodies over the past 22 years. Long-term monitoring of vegetation and the analysis of influential factors are crucial components of studying global changes (Jin et al. 2020). The study findings indicate a significant decrease in NDVI datasets over time (Figure 2). The NDVI values ranged between 0 and 0.7 and were compared with higher and lower values in multi-temporal datasets. In the Chott Merouane region, 2006 exhibited the most negative NDVI values. Several regions also experienced a significant deterioration in vegetation, with values not exceeding 0.2. Additionally, lower healthy vegetation with higher NDVI values was observed due to lower vegetation density. These findings suggest a history of drought and plant stress in these areas. However, certain regions showed an overall rise in NDVI values, surpassing 0.4. The highest NDVI levels in the Sarno watershed were recorded in 2001 (>0.6), indicating abundant vegetation and healthy green spaces. Nevertheless, these same regions have witnessed a substantial decline in vegetation over the years, potentially due to factors such as drought and climatic conditions.
The study findings reveal that the EVI vegetation indicators exhibited stronger spatio-temporal correlations in the study area. While NDVI showed higher values compared with EVI and NDWI, it did not demonstrate a significant relationship with LST based on the obtained results. EVI, on the other hand, emerged as the second-best indicator with a more robust interconnection. This could be attributed to the sensitivity of the EVI to the spectral impact of soil composition and moisture in areas with limited vegetation coverage (Lu et al. 2015). In this research, the weak relationship between NDVI and LST could be due to several factors. NDVI primarily captures the greenness of vegetation and may not directly reflect the thermal properties of the land surface. Additionally, other factors such as soil type, moisture content, and land cover composition can influence the association between NDVI and LST.
The MK test was employed to examine trends in the chronological datasets of NDVI, EVI, NDWI, and LST in the 25 wetlands. The p-value emerges as a critical outcome, providing insights into the significance of the data, the presence of a trend, and whether that trend is monotonous (Guo et al. 2018).
The analysis of NDVI, EVI, NDWI, and LST datasets revealed diverse trends due to their distinct values. The MK test demonstrated a significant monotonically decreasing trend in NDVI (p < 0.05). EVI values, ranging from 0.0001 to 0.951, generally mirrored the NDVI and NDWI results in specific wetlands such as Ain Beida, Chott Ech Chergui, Chott Melrhir, Beni Bahdel dam, Cheliff dam, Dahmoni dam, Golea, Gargar dam, Karrada dam, and Ksob dam. In contrast, these areas with decreasing trends are exacerbating water stress, leading to water deficits and reduced vegetation activity.
On the other hand, the NDVI, EVI, and NDWI values were positive in wetlands such as Arzew saline, Boughezoul dam, Chott Zehrez Chergui, Dayet El Ferd, Chott Zehrez Gharbi, Chott El Hodna, Macta Marsh, Great Sebkha, Telamine Lake, Bougara dam, Boughrara dam, Bouhanifia dam, Chorfa dam, Chott Merouane, and Sarno dam, indicating a positive trend across all datasets. To determine whether the trend is monotonous, a p-value less than 0.05 is considered. The MK test for LST showed that in all wetland regions with a positive monotonous trend, there was a consistently ascending pattern, indicating an undiminished trend (Royden & Fitzpatrick 2010). The presence of a non-monotonous trend suggests fluctuations of both increasing and decreasing trends over the years, indicating variable responses to drought conditions and higher water deficits in different areas. In contrast, the remaining wetland regions displayed a consistent, monotonous trend throughout the entire period.
Examining Sen's slope on time-series data spanning from 2020 to 2022, with a specific emphasis on natural indicators, yields significant knowledge regarding the environmental changes occurring in various wetlands. The results emphasize notable patterns and alterations in the density of vegetation, moisture level, and surface temperature in different geographical areas. The evident and consistent rising trajectory in vegetation density, as evidenced by NDVI and EVI values, suggests favorable ecological transformations. The reported increase was noted in multiple dams, namely Beni Bahdel dam, Bouhanifia dam, Boughrara dam, and Gargar dam. The increase in vegetation density suggests enhanced environmental conditions and may have favorable consequences for biodiversity and ecological well-being in these regions.
The NDWI research has detected a significant rise in the water content of Ksob dam, Boughezoul dam, and Chott Zehrez Gharbi, which is an important observation. This implies fluctuations in water accessibility, which could possibly affect aquatic habitats and the surrounding environment. It is crucial to monitor these changes to comprehend the water dynamics and guarantee sustainable management of water resources.
The progressive increase in surface temperature over a period of time, suggesting climatic circumstances typified by an excessive amount of heat, is a worrisome pattern. High temperatures can negatively impact the density of vegetation, perhaps resulting in decreased biodiversity and modified ecosystems. This discovery highlights the significance of tackling climate-related difficulties and executing plans for climate resilience.
The areas with the highest impact on vegetation cover, namely Dayet El Ferd, Sarno dam, Chott El Hodna, Ain Beida, Arzew saline, Chott Ech Chergui, Telamine Lake, Chott Merouane, Golea, Gargar dam, Karrada dam, and Sarno dam, have been identified as key areas for targeted conservation and environmental management efforts. Gaining a comprehensive understanding of the particular obstacles encountered by these locations is essential for adopting efficient mitigation strategies.
The statistical significance of the trend test results, with a significance threshold of 1%, enhances the reliability of the findings. The continued presence of statistically significant trends, with a confidence level of 95%, from 2020 to 2022 highlights the strength and reliability of the observed changes. This knowledge holds significant value for policymakers and researchers as it enables them to make well-informed decisions on environmental conservation and resource management techniques.
Ultimately, the Sen's slope analysis of time-series data has revealed significant patterns and fluctuations in environmental indicators across different dams and locations. These findings establish a basis for focused environmental interventions, highlighting the importance of sustainable practices and adaptable techniques to tackle the difficulties presented by shifting environmental conditions. Continual surveillance and investigation will be crucial to assess the enduring effects and improve conservation strategies accordingly.
CONCLUSIONS
The comprehensive analysis conducted in this study utilizes an extensive dataset of NDVI spanning from 2000 to 2022 to examine changes in 25 different regions over a period of 22 years. The average mean NDVI values varied from 0 to 0.7, with a mean percentage between 0.098% and 0.34%. The analysis revealed a noticeable decline in NDVI values over the course of the multi-year cycle. Additionally, many regions exhibited significant deterioration in vegetation, with values not exceeding 0.2.
The study found that NDVI had a stronger spatial and temporal response compared with EVI. The correlation between NDVI and EVI was also notable, contributing to the overall findings. The negative association between NDVI and LST underscores the impact of drought and plant stress on vegetation in the study areas (R2 = 0.109 to R2 = 0.5701). The significant negative correlations between NDVI and LST values indicate that higher temperatures are associated with reduced vegetation density and health.
The findings derived from the MK trend analysis underscore the importance of NDVI, highlighting its strong association with EVI. Overall, the findings of this study emphasize the severity of the changes and degradation observed in these areas, which can be attributed to factors such as droughts and other climatic conditions that induce stress on vegetation.
These findings provide valuable information for understanding the dynamics of vegetation and water stress, highlighting the need for effective management and conservation measures in the studied regions.
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.