Land cover transformations in northern Algeria from 2001 to 2021: A MODIS-based spatiotemporal analysis

Bilel Zerouali ¹; Richarde Marques da Silva ²; Nadjem Bailek ³; Valeriano Carneiro de Lima Silva ⁴; Celso Augusto Guimarães Santos ⁴

1, Laboratory of Architecture, Cities and Environment, Faculty of Civil Engineering and Architecture, Department of Hydraulic, Hassiba Benbouali University of Chlef, B.P. 78C, Ouled Fares 02180, Algeria

2, Department of Geosciences, Federal University of Paraíba, João Pessoa, Paraíba, Brazil

3, Laboratory of Mathematics Modeling and Applications, Department of Mathematics and Computer Science, Faculty of Sciences and Technology, Ahmed Draia University of Adrar, Adrar 01000, Algeria

4, Department of Civil and Environmental Engineering, Federal University of Paraíba, João Pessoa, 58051-900, Paraíba, Brazil

 

Manuscript link
http://doi.org/10.30493/DAS.2026.011109

Abstract

Land use/land cover (LULC) changes in northern Algeria from 2001 to 2021 were investigated in this research. The assessment utilized the MODIS MCD12Q1 dataset combined with Geographic Information System (GIS) techniques. Substantial landscape transformations were highlighted by the findings. Forests (+47.52%), shrublands (+17.65%), wetlands (+58.64%), and natural vegetation (+465.72%) experienced notable growth, whereas a drop of −14.13% was documented for barren lands by the end of study period. Moderate overall growth was displayed by agricultural lands, despite various fluctuations throughout the timeline. Concurrently, a continuous expansion (+3.90%) driven by escalating anthropogenic pressure was recorded in urban and built-up areas. Regional landscape evolution is clearly shaped by a combination of factors. Climatic variability, ecological succession, land management practices, and human activities collectively drive these observed dynamics. Partial alignment with previously documented vegetation recovery and urban expansion trends in Algeria was demonstrated by comparative literature reviews. However, the moderate spatial resolution (500 m) of the MODIS MCD12Q1 product likely caused several identified data discrepancies. Ultimately, the immense value of remote sensing and GIS approaches for regional environmental monitoring is proven by this project. Achieving better classification accuracy in heterogeneous Mediterranean environments, however, strictly necessitates the use of higher-resolution datasets.

Keywords: Land cover, MODIS, Vegetation recovery, Northern Algeria

Introduction

Land use and land cover (LULC) dynamics are primary drivers of regional and global environmental change [1][2]. These patterns also highlight how human societies interact with nature. Phenomena like urbanization, agricultural intensification, and ecosystem degradation are frequently driven by land cover alterations. Additionally, strong links exist between these shifts and deforestation, biodiversity loss, or climate variability [3][4]. Recent decades have witnessed an acceleration of these transformations across Northern Algeria and other Mediterranean and semi-arid zones. Rapid population surges and urban expansion largely fuel this rapid pace. Industrial development, intense agricultural pressure, and mounting climate-related stresses play major roles as well [5-7]. Consequently, sustainable environmental management requires a thorough understanding of the spatial and temporal evolution of LULC. Territorial planning, natural resource conservation, and climate adaptation strategies also depend heavily on this spatial knowledge [8][9].

In Algeria, the highest demographic density and economic output of the nation are concentrated within northern region. Highly heterogeneous landscapes define this territory. Coastal plains, mountainous forest ecosystems, and agricultural zones are mixed with urban agglomerations and semi-arid transitional environments [10-12]. Severe anthropogenic and environmental pressures have impacted the area over recent decades. These burdens involve rapid urban sprawl, agricultural land transformation, and recurrent wildfires. Moreover, deforestation, land degradation, and desertification processes further complicate the issue [13][14]. Hydrological functioning, ecosystem services, and food security are heavily affected by these modifications. Furthermore, biodiversity conservation and regional climate dynamics face serious implications. Therefore, a major scientific and environmental priority has been established around the accurate monitoring and analysis of LULC evolution across Northern Algeria [14].

Continuous, large-scale, and economical observations of the planetary surface are supplied by remote sensing and Geographic Information Systems (GIS). These technologies have thus emerged as fundamental tools for tracking long-term LULC changes. Regional and continental-scale LULC studies frequently employ the MODIS Land Cover Type Product MCD12Q1 dataset. Researchers favor this specific archive for its annual temporal resolution and global coverage. A standardized classification scheme and long-term consistency also boost its utility [15-17]. Yearly land-cover maps featuring a 500 m spatial resolution are generated by the MCD12Q1 product, basically relying on the International Geosphere–Biosphere Programme (IGBP) classification system. Consequently, the identification and tracking of major land-cover classes are greatly simplified. Such categories include forests, croplands, urban areas, shrublands, grasslands, wetlands, and barren lands. Long-term environmental changes and land-cover transitions are seamlessly analyzed using this resource, due to its temporal continuity since 2001 [18][19].

LULC dynamics throughout Algeria have been explored by several previous studies. These efforts relied upon varied remote sensing datasets and diverse methodologies [13][20]. However, localized terrains, individual watershed scales, or specific cities have dominated the focus of existing literature. However, comprehensive large-scale assessments spanning multi-decadal periods across the entirety of Northern Algeria remain quite scarce. Additionally, while descriptive mapping was primarily prioritized by many past researchers, spatial patterns of land-cover transitions and urban expansion processes were rarely investigated in depth. Vegetation degradation trends and the environmental implications linked to these shifts were similarly overlooked. Therefore, an integrated regional-scale analysis is still required to assess the dominant trajectories of LULC transformation and their spatial heterogeneity across Northern Algeria.

Given this background, the spatiotemporal dynamics of LULC within Northern Algeria are analyzed in the current investigation. Annual MCD12Q1 land-cover data serve as the foundation for this work. The evolution of major land-cover classes throughout the study period is systematically examined. Furthermore, dominant transition pathways connecting distinct categories are pinpointed. The spatial distribution of regional environmental transformations is also evaluated. Specific focus is directed toward urban expansion, agricultural dynamics, and vegetation changes. The steady advance of barren or sparsely vegetated areas is heavily scrutinized as well. A long-term regional-scale LULC assessment forms the core originality of this work. Consistent annual MODIS observations are merged with spatial and temporal analyses to map land-cover transitions across Northern Algeria and provide an updated understanding of recent environmental transformations throughout the region.

Materials and Methods

The study area

The study area is located in Northern Algeria, covering approximately 332,769 km². According to the National Agency of Water Resources (ANRH), the region is divided into 17 major river basins (Fig. 1), and bordered by the Mediterranean Sea to the north, Morocco to the west, Tunisia to the east, and the Saharan Atlas to the south. The topography is mainly structured by two east–west mountain ranges: the Tell Atlas in the north and the Saharan Atlas in the south.

Northern Algeria is characterized by a Mediterranean climate with mild wet winters and hot dry summers, while the interior plateaus experience colder winters and lower rainfall. Annual precipitation generally ranges from about 1000 mm in the north to nearly 1500 mm in the northeastern regions. The northeastern part of the country hosts the most productive forest ecosystems under humid and sub-humid climatic conditions, particularly in regions such as Kabylie, El Kala, and Souk-Ahras.

Nevertheless, severe degradation has heavily impacted these natural environments. Recurrent wildfires, deforestation, and overgrazing constantly drive this ecological decline. Flawed agricultural practices and mounting human demands further accelerate the habitat damage. Substantial land-use and land-cover shifts throughout Northern Algeria were documented by previous remote-sensing studies. Massive physical alterations across forest, agricultural, and urban landscapes were tracked by these earlier investigations, highlighting the escalating vulnerability of the territory to anthropogenic stress and environmental deterioration [14][21][22].

MODIS land-cover data

Land-use and land-cover changes were investigated using the Terra and Aqua combined Moderate Resolution Imaging Spectroradiometer Land Cover Type product, MCD12Q1 Version 6.1. The product provides global land-cover information annually at a nominal spatial resolution of 500 m. This data is derived from supervised classification of spectrotemporal information obtained from MODIS Terra and Aqua observations, followed by post-processing using ancillary information and prior knowledge.

The dataset was accessed through the Google Earth Engine data catalogue using the image-collection identifier MODIS/061/MCD12Q1. The analysis covered 2001–2021 and therefore included 21 annual MCD12Q1 images with a spatial resolution of 500 m. All available annual images were processed to examine temporal consistency, whereas 2001, 2010, and 2021 were selected as benchmark years for detailed mapping and interval-based change assessment. These years represent the beginning, approximate midpoint, and end of the study period and allow comparison of the intervals 2001–2010, 2010–2021, and 2001–2021.

The LC_Type1 band was selected since it represents the International Geosphere–Biosphere Programme (IGBP) classification scheme. The IGBP legend contains 17 land-cover classes, numbered from 1 to 17, in addition to the fill or unclassified value. The QC band was also inspected to identify classification-quality conditions and unclassified pixels. Fill and unclassified pixels were excluded from the quantitative calculations. The product and original IGBP class definitions were obtained from the official MCD12Q1 Version 6.1 User Guide [23].

Data preprocessing

A consistent preprocessing procedure was applied to all images to prevent differences in extent, grid alignment, projection, or class coding from being interpreted as land-cover changes.

Temporal filtering and band selection

The MCD12Q1 collection was filtered by year from 1 January 2001 to 31 December 2021. For every year, the corresponding annual image was selected, and the LC_Type1 and QC bands were extracted. The number and dates of the returned images were checked to confirm that one valid annual land-cover image was available for each year.

Because MCD12Q1 is an annual land-cover product, no monthly compositing, atmospheric correction, cloud masking, or spectral-index calculation was required. Such operations are already incorporated into the production of the MODIS land-cover product.

Spatial masking and clipping

The study-area polygon was uploaded to Google Earth Engine and used to clip every annual image to the same geographical extent. Pixels outside the study-area boundary were masked. Unclassified and fill pixels were also masked before area calculations.

All annual rasters were maintained on the native MCD12Q1 grid during pixel-based comparison. This process allowed avoiding spatial displacement and artificial land-cover transitions that could result from independently resampling each annual raster.

Projection and resampling

The original MCD12Q1 data are distributed in the MODIS sinusoidal projection. Quantitative area calculations were performed using the native raster geometry and the Google Earth Engine ee.Image.pixelArea() function. This function calculates the ground area of each valid pixel and avoids assuming that every nominal 500 m pixel represents exactly the same area after reprojection.

For visualization and map preparation, the resulting images were exported and displayed in a common geographic or projected coordinate system. Nearest-neighbor resampling was used whenever categorical land-cover rasters were reprojected. Bilinear and cubic interpolation were not used since they can generate invalid intermediate class values and modify class boundaries.

Quality-control treatment

The MCD12Q1 QC layer was inspected to identify the general classification status of the pixels. Pixels carrying the unclassified or fill value were excluded. The original land-cover labels of classified pixels were retained to avoid changing the official MCD12Q1 classification.

The quality layer was used as a product-level diagnostic rather than as an independent accuracy assessment. No claim of local classification accuracy was made because independent field observations or a consistent high-resolution reference dataset covering the entire region and all study years were unavailable.

Land-cover classification and reclassification

The input images were already classified in the MCD12Q1 product. Therefore, the classification stage of the present analysis consisted of selecting the IGBP LC_Type1 layer and aggregating its original classes into broader categories appropriate for regional-scale change assessment.

The 17 original IGBP classes were reclassified into eight main categories:

1. Forests: IGBP classes 1–5.
2. Shrublands: classes 6 and 7.
3. Savannas and grasslands: classes 8–10.
4. Permanent wetlands: class 11.
5. Agricultural lands: classes 12 and 14.
6. Urban and built-up areas: class 13.
7. Barren or sparsely vegetated lands: class 16.
8. Water and ice: classes 15 and 17.

The reclassification can be expressed as:

R(p)=f[C(p)]

where C(p) is the original IGBP label of pixel p, f is the reclassification rule, and R(p) is the resulting aggregated land-cover category.

The same reclassification lookup table was applied without modification to all 21 annual images. This ensured that apparent temporal changes were not caused by differences in class aggregation among years.

Calculation of land-cover areas

The area of each reclassified land-cover category was calculated for every annual image. For a given class i and year t, a binary indicator raster was created:

The area occupied by class i in year t was then calculated as:

where A_i,t is the area of class i in square kilometers, N is the total number of pixels within the study area, and a_p is the ground area of pixel p in square meters. Division by 10⁶ converts square meters to square kilometers.

The proportional coverage of each class was calculated as:

where P_i,t is the percentage of the study area occupied by class i, and A_T is the total valid mapped area.

A closure test was conducted for every year by comparing the sum of the individual class areas with the total valid mapped area:

where k=8 is the number of aggregated classes. A value close to zero confirms that all valid pixels were assigned only once.

Land-cover change detection

A post-classification comparison approach was used to quantify land-cover changes. This method compares independently labelled categorical maps and is suitable for an already classified product such as MCD12Q1. Change was evaluated for 2001–2010, 2010–2021, and 2001–2021. The analysis included class-area differences, proportional changes, gains and losses, spatial change maps, and pixel-level transition matrices.

Absolute area change

The absolute change in class i between years t₁ and t₂ was calculated as:

A positive value indicates a net gain, whereas a negative value represents a net loss.

Relative percentage change

The relative change was calculated as:

This statistic expresses the change relative to the initial area of the class. Percentage changes were interpreted cautiously for classes with small initial areas because relatively small absolute changes can generate large percentage values.

Mean annual rate of change

The mean annual net change was calculated as:

where M_i is expressed in square kilometres per year (km²/year).

The compound annual rate of change was calculated as:

where R_i is expressed as a percentage per year.

Temporal analysis

The area of every aggregated land-cover category was extracted for each year from 2001 to 2021. The resulting 21-year series was used to examine whether changes between the benchmark years represented persistent tendencies or short-term fluctuations. For each class, the temporal series was represented as:

{A_i,2001, A_i,2002,…,A_i,2021}

Annual differences were calculated as:

D_i,t = A_i,t – A_i,t-1 

Large year-to-year changes were inspected because categorical global products can exhibit temporal label instability, particularly in heterogeneous landscapes and transition zones. Consequently, a change observed between only two individual years was not automatically interpreted as a permanent ecological transformation. Interpretation considered the complete temporal trajectory and the known limitations of the 500 m MCD12Q1 product.

Spatial interpretation

The reclassified maps and transition products were examined to determine the spatial distribution of land-cover persistence, gain, loss, and conversion. Particular attention was given to forest changes in northern mountainous areas; agricultural dynamics in coastal and inland plains; urban expansion around major settlements; shrubland and grassland changes in semi-arid transition zones; changes between barren and vegetated categories; and wetland and water-body variations.

Spatial patterns were interpreted in relation to the regional environmental setting. However, causal attribution was avoided because the MCD12Q1 maps identify land-cover changes but do not directly determine whether they resulted from climate variability, fire, agricultural management, urbanization, afforestation, or other drivers.

Software and computational platforms

Google Earth Engine was used for cloud-based access to MCD12Q1 Version 6.1, temporal filtering, band selection, clipping, quality-control inspection, reclassification, pixel-area calculation, annual class-area extraction, raster overlay, transition analysis, and export of the results.

ArcGIS 10.2.2 was used for the preparation of the study-area boundary, geometry verification, cartographic visualization, map-layout production, and final inspection of exported raster and vector layers. Nearest-neighbor resampling was specified for categorical rasters.

Tabular outputs were exported in CSV format and organized in Microsoft Excel for independent checking, summary-table preparation, and figure-data verification. Statistical calculations followed the equations provided above.

Methodological quality control and limitations

Several checks were performed to reduce processing errors. All annual maps were clipped using the same boundary, retained on a common raster grid, and reclassified using the same lookup table. Categorical layers were not processed using continuous-data interpolation. Class-area closure was checked for every year, and the row and column totals of each transition matrix were compared with the corresponding class areas at the beginning and end of each interval.

Nevertheless, the analysis inherits the limitations of MCD12Q1. Its nominal 500 m resolution may not adequately represent fragmented urban areas, narrow agricultural fields, small wetlands, or heterogeneous forest–shrubland mosaics. Mixed pixels and temporal label instability may produce apparent transitions that do not represent actual land-cover conversion. Furthermore, no independent regional accuracy assessment was possible because consistent reference observations were unavailable for all years. The results should therefore be interpreted as regional-scale land-cover patterns rather than parcel-level changes.

Results

Spatial distribution of LULC classes

Noticeable spatial and temporal changes across Northern Algeria between 2001, 2010, and 2021 were exposed by the LULC maps (Fig. 2). Barren lands maintained strict dominance across the southern margins and deep inland territories throughout the investigation. The severe environmental constraints of these arid zones are highlighted by this persistence. Conversely, croplands and shrublands clustered predominantly along the northern coastal belts. The adjacent sub-humid zones were also encompassed by these fertile corridors. A stark geographical divide is consequently illustrated by the distribution of these specific land-cover categories. The immediate coastal proximity clearly fosters denser vegetation networks compared to the desolate interior.

Data spanning 2001 to 2010 revealed a distinct expansion across shrublands, agricultural lands, and savannas and grasslands. A concurrent reduction in barren areas accompanied this widespread vegetation areas growth. Favorable environmental conditions likely supported the broad proliferation of these diverse botanical classes. Moderate spatial gains were also achieved by regional forest cover. This localized woodland expansion materialized primarily within the rugged northeastern mountainous sectors. However, a notable reversal characterized the subsequent 2010 to 2021 timeframe. Evident regression afflicted agricultural lands alongside the sprawling savanna and grassland ecosystems. Concurrently, a partial resurgence of barren lands was documented across numerous inland tracts. A period of severe ecological stress or rapidly shifting land management practices is strongly suggested by this latter dynamic.

Prominent coastal metropolitan centers retained the highest concentrations of urban and built-up lands. A gradual outward expansion of these developed surfaces materialized consistently over time. The relentless pace of regional urbanization and infrastructural sprawl is underscored by this continuous spatial growth. Meanwhile, relatively restricted spatial variability was exhibited by wetlands and water bodies. A highly stable physical footprint was maintained by these critical hydrological features for the duration of the assessment. Ultimately, profound landscape heterogeneity is indicated by these comprehensive LULC maps. A complex intersection of variables drives the evolving land-cover patterns across Northern Algeria. Fluctuating climatic conditions and underlying topography establish the fundamental environmental baseline. Simultaneously, shifting vegetation dynamics and escalating anthropogenic pressures exert immense influence over this ongoing terrestrial transformation.

Land cover distribution and temporal evolution

Significant land-cover changes across Northern Algeria between 2001 and 2021 were revealed by the analysis of the MODIS Land Cover Type Product MCD12Q1 data (Fig. 3). The dominant category throughout this entire period remained barren lands. Initially, a massive 141,507.95 km² (42.52%) was covered by this class in 2001. This footprint then was decreased to 105,886.00 km² (31.82%) by 2010. Eventually, an expansion to 121,512.93 km² (36.52%) was recorded in 2021.

The second-largest landscape category was formed by shrublands, with a steady expansion from 82,480.49 km² (24.79%) in 2001 to 97,036.96 km² (29.16%) in 2021 recorded under this category. Simultaneously, a substantial fraction of the territory was occupied by agricultural lands. These cultivated tracts grew from 52,869.92 km² (15.89%) in 2001 to reach 63,415.94 km² (19.06%) in 2010. However, a subsequent decline to 55,456.50 km² (16.67%) was documented by 2021. A similar trajectory was followed by savannas and grasslands. Their coverage initially increased from 49,453.65 km² (14.86%) in 2001 to 59,366.56 km² (17.84%) by the end of the first decade. Afterward, a reduction to 50,573.32 km² (15.20%) was observed in 2021.

A moderate increase between 2001 and 2010 was exhibited by forest areas. Specifically, wooded regions expanded from 2816.56 km² (0.85%) in 2001 to 4540.78 km² (1.36%) in 2010. A minor contraction to 4155.03 km² (1.25%) followed this peak in 2021. Meanwhile, a gradual upward trend was charted by urban and built-up lands. These artificial surfaces grew from 2635.12 km² (0.79%) in 2001 to 2737.84 km² (0.82%) in 2021. The continuous urbanization processes sweeping across the territory are clearly reflected by this steady growth.

Relatively minimal portions of the total landscape were occupied by permanent wetlands, natural vegetation, and water bodies. In spite of this small footprint, an increase from 126.33 km² (0.04%) in 2001 to 200.41 km² (0.06%) in 2021 was achieved by wetlands. Concurrently, natural vegetation expanded significantly, surging from 46.88 km² (0.01%) to 265.21 km² (0.08%). High stability was maintained by water bodies throughout the entire investigation. These aquatic features consistently fluctuated around just 0.25% of the overall surface area.

Land cover change dynamics

Substantial shifts in land-cover dynamics across Northern Algeria between 2001 and 2021 were uncovered by the temporal analysis (Table 1). The largest overall growth throughout the assessment was recorded by shrublands. These zones expanded by 14,556.48 km² over the twenty-year span. The majority of this addition took place from 2001 to 2010 (+13,245.65 km²). A more restrained expansion (+1310.83 km²) followed in the 2010–2021 timeframe. A net gain of 1338.47 km² was also documented in forest areas over the total studied duration, which represented a 47.5% addition to this category based on its initial area.

Contrasting dynamics were experienced by agricultural lands and savannas/grasslands. Agricultural lands grew noticeably from 2001 to 2010 (+10,546.01 km²). However, a subsequent reduction (−7959.43 km²) during the latter years produced a modest final addition of 2586.58 km² to the initial agricultural area. An analogous pattern was observed in savannas and grasslands. These ecosystems expanded by 9912.91 km² in the initial ten years. This early surge preceded a severe drop (−8793.24 km²) after 2010.

A continuous rise was exhibited by urban and built-up lands throughout the assessment. Ongoing regional urbanization processes were reflected by this cumulative expansion of 102.71 km². Progressive growth was similarly recorded in permanent wetlands and natural vegetation. Nevertheless, the total spatial extent of these specific categories stayed fairly constrained.

Conversely, the most pronounced reduction from 2001 to 2010 was experienced by barren lands (−35,621.95 km²). A partial rebound (+15,626.93 km²) occurred in these arid sections during the subsequent 2010–2021 interval. Despite this observation, a negative overarching trajectory (-14.13%) was maintained between 2001 and 2021, resulting in a cumulative loss of 19,995.02 km² for this category. General stability was maintained by water bodies. These aquatic features displayed merely trivial fluctuations across the entire timeframe.

Discussion

The LULC analysis revealed significant environmental changes between 2001 and 2021. Vegetation-related classes such as forests (+47.52%), shrublands (+17.65%), natural vegetation (+465.72%), and wetlands (+58.64%) generally increased, indicating progressive vegetation recovery and ecological expansion in several areas. In contrast, barren land showed an overall decrease (−14.13%), suggesting partial conversion of degraded surfaces into vegetated or agricultural lands. Agricultural lands exhibited moderate long-term growth (+4.89%), although a decline was observed during 2010–2021, possibly linked to land degradation, climatic stress, or urban pressure. Urban and built-up areas continuously expanded (+3.90%), reflecting ongoing human development and infrastructure growth. Overall, the results highlight the combined influence of climate variability, land management practices, ecological succession, and anthropogenic activities on landscape dynamics across the study region.

The observed LULC dynamics generally agree with several previous studies conducted in Algeria, particularly regarding the continuous environmental transformation driven by anthropogenic pressure, climatic variability, and land management practices. Researchers reported important fluctuations in vegetation cover in the Tlemcen region, characterized by forest degradation due to fires, followed by partial vegetation recovery [24]. Similarly, the significant vegetation recovery in the Chréa Protected Area during earlier decades was reported in another study [25], supporting the vegetation increase observed in the present study.

The increase in shrublands and natural vegetation observed here is also partially consistent with other studies, where a progressive expansion of dense vegetation in the Tolga region associated with agricultural growth and ecological changes were identified [26]. In contrast, several studies reported substantial forest decline caused by repeated fires and land degradation. For instance, a study documented a major reduction in forest cover in Sidi Bel Abbes [27], while another associated increases in barren lands and degraded forests with enhanced soil erosion processes in northeastern Algeria [28].

Urban growth identified in this study also agrees with previous literature [22][29][30], all reported continuous urban expansion at the expense of agricultural and natural lands in northern Algeria. However, the relatively limited urban increase (+3.90%) and the strong vegetation recovery detected in the present work differ from some regional studies that showed more pronounced urbanization and vegetation degradation trends.

These discrepancies may be related to the characteristics of the MODIS MCD12Q1 product used in this study. Although MCD12Q1 is highly useful for large-scale temporal monitoring, its 500 m spatial resolution may not provide sufficient accuracy for heterogeneous Mediterranean landscapes such as northern Algeria. Mixed pixels, confusion between sparse vegetation and barren lands, and difficulties in distinguishing fragmented urban and agricultural areas can introduce classification uncertainty. Consequently, MCD12Q1 may overestimate vegetation recovery or underestimate urban expansion and land degradation in complex environments. However, the presented results can provide a general idea of the trends in LULC changes in northern Algeria during the period 2000-2021, which might assess the direction of future localized studies within the studied region. Therefore, higher-resolution datasets such as Landsat or Sentinel imagery should be utilized for more reliable local scale assessments.

Conclusions

Substantial landscape transformations in northern Algeria between 2001 and 2021 were uncovered by the LULC analysis presented in this work. Forests, shrublands, wetlands, and natural vegetation expanded during this timeframe. Simultaneously, barren lands experienced a broad decline. Ongoing anthropogenic pressure and land-use conversion processes drove a continuous expansion of urban and built-up areas. Several constraints limit the scope of these findings. Classification accuracy in heterogeneous Mediterranean environments like northern Algeria might be compromised by the 500 m spatial resolution of the MODIS MCD12Q1 dataset. Thus, higher-resolution satellite imagery, including Landsat or Sentinel data, should be integrated into upcoming localized investigations in order to elevate classification precision and enhance local-scale interpretation. The fundamental drivers of LULC dynamics could be clarified by incorporating field observations alongside machine learning classification techniques. Moreover, adding climate-related variables would provide necessary contextual depth to these models. Subsequent assessments should evaluate the correlation between LULC changes and specific environmental impacts across northern Algeria.

Conflict of interest statement
The authors declared no conflict of interest.
Funding statement
The authors declared that no funding was received in relation to this manuscript.
Data availability statement
The authors declared that all related data and developed code will be available upon reasonable request from the corresponding author.

References

1. Akpatsu IB, Abera W, Chehbouni A, Laamrani A. Land Use and Land Cover Transitions in Mountainous Landscapes: A Systematic Review of Soil Carbon Dynamics, Challenges, and Research Perspectives. Environments. 2026;13(5):269. DOI
2. Li X, Chen G, Liu X, Liang X, Wang S, Chen Y, Pei F, Xu X. A New Global Land-Use and Land-Cover Change Product at a 1-km Resolution for 2010 to 2100 Based on Human–Environment Interactions. Ann. Am. Assoc. Geogr. 2017;107(5):1040-59. DOI
3. Kashyap A, Parashar D, Bera D, Kumar A, Sharma A. Assessing the Impact of Habitat Changes on Ecosystem Regions Using Vegetation Indices: A Case Study of Land-Use Dynamics in Namchi, Sikkim. In: Sustainability Solutions. Springer Nature Switzerland. 2026. DOI
4. Seeboonruang U, Mandadi R, Thammaboribal P, Gonzales AL, Ganni SVSAB. Land Use Classification, Prediction, and the Relationship Between Land Use and Sediment Loss in the Lam Phra Phlong Watershed, Thailand. Agriculture. 2026;16(4):448. DOI
5. John CK, Pu JH. Climate-Driven Water Scarcity and Its Public Health Implications: A Multi-Regional Assessment Across Vulnerable Socio-Ecological Systems. Water. 2026;18(6):699. DOI
6. Anam S, Umar A, Fatima I, Majeed M, Fatima S, Naz Z, Khan UA, Khalil M, Khattak WA, Anas M. Climate Change and Innovative Solutions for Drought and Desertification. In: Sustainability Solutions. Springer Nature Switzerland. 2026. DOI
7. Mathbout S, Boustras G, Papazoglou P, Martin‐Vide J, Lopez‐Bustins JA. Climate Crisis, Human Mobility and Security Challenges in the MENA Region: Implications for Sustainable Development and Regional Stability. Sustain. Dev. 2026:sd.70820. DOI
8. Miah MT, Fariha JN, Jodder PK, Hasan MR, Hossain MZ, Rahaman KR. Remote Sensing–Based Spatiotemporal Modeling of Land Use Land Cover Changes and Associated Risks in Coastal Ecosystems: Stochastic Approach to Sustainable and Climate-Resilient Urban Development. J. Urban Plann. Dev. 2026;152(2):04026025. DOI
9. Dong Y, Chang X. Spatio-temporal analysis of land use and land cover changes: assessing urban expansion and ecological transformations (2003–2023). GeoJournal. 2026;91(2):40. DOI
10. Ballah A, Chahid M, Aqnouy M, Elhoussaoui A, Lahcen B, Alrasheedi KG, Afandi GE. Using the WEAP Model to Analyze Water Supply and Demand Dynamics in the Tamanghasset Watershed Under Future Stressors. Earth Syst. Environ. 2026;10(8):9995-10012. DOI
11. Islam TM. Strategic Transformation of Algeria into a Competitive Tourism Destination: An Integrated Analysis of Marketing Policies, Spatial Development, and Tourism Hubs within the SDAT 2030 Framework. Sci. Educ. Innov. Mod. Probl. 2026;9(4):1-16. DOI
12. Bouregaa T. Climate Change in Algeria: Impacts, Vulnerabilities, and Strategic Responses for Sustainable Development. In: Extreme Climate Events, Loss and Damage in Africa. Springer Nature Switzerland. 2026. DOI
13. Zerouali B, Ayek AAE, Bailek N, Kuriqi A, Wong YJ, Srivastava A, Islam ARMT, Ouadja A, Santos CAG. RUSLE model insights for soil conservation and sustainable land use in semiarid environments. Euro-Mediterr. J. Environ. Integr. 2024;10(2):853-76. DOI
14. Zerouali B, Santos CAG, do Nascimento TVM, Silva RMD. A cloud-integrated GIS for forest cover loss and land use change monitoring using statistical methods and geospatial technology over northern Algeria. J. Environ. Manag. 2023;341:118029. DOI
15. Mashala MJ, Dube T, Mudereri BT, Ayisi KK, Ramudzuli MR. A Systematic Review on Advancements in Remote Sensing for Assessing and Monitoring Land Use and Land Cover Changes Impacts on Surface Water Resources in Semi-Arid Tropical Environments. Remote Sens. 2023;15(16):3926. DOI
16. Dangayach R, Pandey AK. Technologies and Methods for Land Use and Land Cover: A Comprehensive Review. In: Advances in Geographical and Environmental Sciences. Springer Nature Singapore. 2025. DOI
17. Szafarczyk A, Agbasi OE. Emerging trends in gis and remote sensing technologies for environmental monitoring: innovations, applications, and future directions. Geoinform. Pol. 2025;24:25. DOI
18. Zhao J, Dong Y, Zhang M, Huang L. Comparison of identifying land cover tempo-spatial changes using GlobCover and MCD12Q1 global land cover products. Arab. J. Geosci. 2020;13(16):792. DOI
19. Liang D, Zuo Y, Huang L, Zhao J, Teng L, Yang F. Evaluation of the Consistency of MODIS Land Cover Product (MCD12Q1) Based on Chinese 30 m GlobeLand30 Datasets: A Case Study in Anhui Province, China. ISPRS Int. J. Geo-Inf. 2015;4(4):2519-41. DOI
20. Hind M, M’hammed S, Djamal A, Zoubida N. Assessment of land use–land cover changes using GIS, remote sensing, and CA–Markov model: a case study of Algiers, Algeria. Appl. Geomat. 2022;14(4):811-25. DOI
21. Guehaz R, Sivakumar V. A case study about the forest fire occurred on 05 July 2021 over Khenchela province, Algeria, using space-borne remote sensing. Front. Remote Sens. 2023;4:1289963. DOI
22. Smail M, Boubaker Z, Sbabdji M, Mouaissa H, Kombate B. Detecting land use/land cover changes and forest degradation: A case study of the lower Soummam valley, northern Algeria. J. For. Sci. 2024;70(3):122-34. DOI
23. Friedl M, Sulla-Menashe D. MODIS/Terra+ Aqua land cover type yearly L3 global 0.05 Deg CMG V061. NASA EOSDIS Land Processes Distributed Active Archive Center (DAAC) data set. 2022:MCD12C1-061. DOI
24. Bardadi A, Souidi Z, Cohen M, Amara M. Land Use/Land Cover Changes in the Tlemcen Region (Algeria) and Classification of Fragile Areas. Sustainability. 2021;13(14):7761. DOI
25. Mekideche S, Belaroui A. Land use/land cover change trends in the Chrea Protected Area of northern Algeria (1984-2021), using Landsat TM+/OLI_TIRS data. Bois For. Trop. 2026;363:1-20. DOI
26. Ben Ratmia MAE, Fezzai S, Zitouni DC, Matallah ME, Bouzaher S, Ben Ratmia FZ, Ramírez-Guerrero G. Identifying hotspots of urbanization and agricultural growth: a spatio-temporal analysis of land use dynamics in Tolga, Algeria (1994–2024). Arid Land Res. Manag. 2025;39(4):602-24. DOI
27. Mohammed G, Boutkhil M, Ayoub B. Assessing the impact of anthropogenic activities on land use and land cover changes in the semi-arid and arid regions of Algeria. Environ. Monit. Assess. 2024;196(4):383. DOI
28. Bouzeria H, Eddine TS, Hamza B, Oussama D, Saâdia B. Evaluating the Effect of Land Use Land Cover Changes on Soil Loss Distribution in the Seybouse Basin, Northeastern Algeria. Dokl. Earth Sci. 2023;510(1):335-48. DOI
29. Senadi I, Zeroual A, Meddi H, Zhang X, Alkama R. Spatio-temporal analysis of Land Use and Land Cover Changes in the Mitidja plain (1983–2023): impact of urbanization on agricultural land. Model. Earth Syst. Environ. 2025;11(4):295. DOI
30. Mahcer I, Baahmed D, Chemirik CHK, Nedjai R. Mapping Environmental Impacts in North-Western Algeria Through Multivariate Spatio-Temporal Analysis Using Remote Sensing and GIS. Ecol. Eng. Environ. Technol. 2024;25(5):42-60. DOI

Cite this article:

Zerouali B, da Silva RM, Bailek N, de Lima Silva VC, Santos CAG. Land cover transformations in northern Algeria from 2001 to 2021: A MODIS-based spatiotemporal analysis. DYSONA-Applied Science. 2027;8(1):71-83. doi: 10.30493/das.2026.011109