Chain-Spectrum Analysis of Land Use/Cover Change Based on Vector Tracing Method in Northern Oman

Abstract

Land use/cover (LUCC) change in arid oasis–desert ecotones has significant implications for spatial governance in ecologically fragile regions. To better capture the temporal and spatial complexity of land transitions, this study developed a vector tracing method by integrating time-series remote sensing data with vector-based transfer pathways. Analysis of northern Oman from 1995 to 2020 revealed the following: (1) Arable land and impervious surfaces expanded from 0.51% to 1.09% and from 0.31% to 0.98%, respectively, while sand declined from 99.03% to 97.01%. Spatially, arable land was concentrated in piedmont irrigation zones, impervious surfaces near coastal cities, and shrubland and grassland along the Al-Hajar Mountains, forming a complementary land use mosaic. (2) Human activities were the dominant driver, with typical one-way chains accounting for 69.76% of total change. Sand was mainly transformed into arable land (7_C1, 7_D1, 7_E1; where the first part denotes the original type, the letter denotes the year of change, and the last digit denotes the new type), impervious surfaces (7_C6, 7_D6, 7_E6), and shrubland (7_E4). (3) Water scarcity and an arid climate remained primary constraints, manifested in typical reciprocating chains in the oasis–desert interface (7_D1_E7, 7_A1_B7, 7_C1_D7) and in the arid vegetation zone along the Al-Hajar Mountain foothills (7_D3_E7, 7_C3_D7), together accounting for 24.50% of total change. (4) The region exhibited coordinated transitions among oasis, urban, and ecological land, avoiding the common conflict of cropland loss to urbanization. During the study period, transitions among arable land, impervious surfaces, forest, shrubland, and wetland were rare (Type 16: 3.31%, Type 82: 2.89%, Type 12: 0.04%, Type 18: 0.01%). The case of northern Oman provides a valuable reference for collaborative spatial governance in ecologically fragile arid zones. Future research should integrate socio-economic drivers, climate change projections, and higher-temporal-resolution data to enhance the applicability of the chain-spectrum method in other arid regions.

1. Introduction

Land use/cover change (LUCC), as an important manifestation of the spatiotemporal dynamics of surface landscapes [1], has significant impacts on global environmental change, terrestrial ecosystems, and human activities [2,3]. Understanding its processes and patterns is essential for deep insights into LUCC mechanisms and effects, as well as for simulation, prediction, and regulation [4]. With the continuous advancement of remote sensing monitoring and geographic information analysis technologies [5], substantial progress has been achieved in LUCC research [6,7]. This progress is not only reflected in substantial improvements in the timeliness and accuracy of data acquisition [8] but also in the expansion of analytical methods from static quantitative descriptions to dynamic process simulations [9]. Traditional quantitative analysis methods usually start with representing the “quantity” change in LUCC, with commonly used indicators the land use dynamic indices [10], land use intensity [11], and the structural proportion of land use [12]. These indices play an important role in revealing overall rates and scales of change. For example, Yhdego et al. [13] used land use dynamics to study LUCC in the Nile River Basin, Northeast Africa, in 2000–2020. Dehghani et al. [14] applied multi-source spatial data and comparative analysis to assess LUCC in Tehran (Iran) and Sydney (Australia) from 1993 to 2023. Zewdu et al. [15] employed Landsat imagery and meteorological data to investigate interactions between LUCC and climate change from 1985 to 2013 in Raichur District, India. LUCC involves not only quantitative but also directional changes, thus bearing vector properties [16]. However, due to the complexity and nonlinearity of LUCC, traditional index models based on “net area” change [17], although capable of characterizing aggregate changes, have inherent shortcomings in representing spatial replacement processes. Net analysis often obscures bidirectional flows of LUCC, leading to misinterpretation of implicit conversion processes (e.g., oasis–desert feedback), thereby hindering a deeper understanding of land system evolution mechanisms. To overcome these limitations, Ma et al. [18] introduced the concept of “flow” from dynamic material changes and proposed a LUCC flow model based on a vector perspective in 2013. This approach traced bidirectional fluxes between LUCC types, enabling refined recording of directions, intensities, and pathways of changes, thus providing a new paradigm beyond traditional methods. In 2017, the same team further proposed the concept of LUCC transfer chains [19], and discovered distinct bidirectional LUCC chains in the oasis–desert ecotone of arid regions, which revealed the long-term and arduous nature of desertification control. The introduction of LUCC chains provided a novel perspective for studying dynamic trajectory characteristics of LUCC in time series. However, current research remains limited by the disconnection between temporal sequence analysis and spatial atlas analysis, which restricts accurate identification of when and where key LUCC processes occur (time, location, and extent). Thus, it is urgently necessary to integrate LUCC change chains with characteristic spatial atlases, in order to provide methodological support for identifying critical areas of land system sustainability.

Arid and semi-arid regions account for about 40% of the Earth’s land surface, where the fragility of the water-soil-atmosphere system makes LUCC a sensitive indicator of global change [20,21]. The arid oasis–desert ecotone in West Asia, as a typical ecological transition zone, reflects the dynamic interplay between human activities and the natural environment [22]. This region is severely affected by vegetation degradation and intensified desertification [23], posing threats to ecological security and sustainable development. Oman, as a typical ecologically fragile arid region [24], has received relatively little attention in LUCC studies. The LUCC evolution process in Oman embodies the dynamic trade-offs between urbanization, agricultural development, and ecological restoration in arid countries, and thus is representative for understanding development models in such regions. This study innovatively constructs a spatio-temporal coupling characterization method of LUCC based on vector attributes, integrates multi-temporal remote sensing monitoring to quantitatively analyze the spatio-temporal heterogeneity of LUCC in northern Oman, and focuses on overcoming the disjunction between temporal and spatial analyses. The aim is to address the traditional separation of “process-pattern” analysis and to provide methodological and decision-making support for the sustainable development of LUCC systems in the oasis–desert ecotone of arid regions. The objectives of this study were to: (1) identify transfer chains of LUCC in northern Oman with a particular focus on dominant one-way and reciprocating chains; (2) construct a spatial atlas of LUCC change through chain classification and spatial distribution mapping; (3) apply a vector tracing method to characterize the spatio-temporal evolution of LUCC and establish the linkage between attribute dimensions and spatial patterns thereby proposing support for governance.

2. Materials and Methods

2.1. Study Area

Oman is located in the southeastern part of the Arabian Peninsula, bordering the Arabian Sea and the Gulf of Oman, and lies at the geopolitical crossroads of Asia and Africa (Figure 1). It exhibits typical arid-zone geographical characteristics, with the northern region being the most topographically complex, richest in water resources, and most agriculturally developed. The Al-Hajar Mountains extend from the Musandam in the northeast to the Sharqiyah region in the southeast, flanked by the Rub’ al Khali Desert to the southwest and the Wahiba Sands to the southeast. The terrain alternates among mountains, alluvial fans, and coastal plains such as Al-Batinah and Sharqiyah, forming a distinctive geomorphology-hydrology-climate gradient system. Annual precipitation in the high mountains can reach about 300 mm, compared to 100 mm in piedmont and coastal zones. Traditional falaj systems [25] channel groundwater for irrigation, sustaining oases and agricultural zones that cultivate drought-resistant crops such as dates, grains, and vegetables. In 2020, sand was the dominant type (97.01%), while arable land and impervious surface accounted for 1.10% and 0.98% of the area, other LUCC types are summarized (

Table 1. Summary of Land Use/Cover Change (LUCC) statistics for northern Oman in 2020.

Type	Area/km2	Proportion
Arable land	1384.95	1.10
Forest	196.53	0.15
Grassland	336.10	0.26
Shrubland	477.91	0.38
Water body	1250.02	0.02
Impervious surfaces	29.39	0.98
Sand	123,166.88	97.01
Wetland	120.71	0.10

). This pronounced topographic and climatic gradient, combined with seasonal variability, makes the region highly sensitive to climate change, as recent decades have seen decreasing annual rainfall and rising mean temperatures. These shifts have intensified water scarcity and altered vegetation growth cycles, thereby influencing agricultural expansion, urban development, and patterns of land transformation.

2.2. Data Sources and Processing

The primary datasets used in this study include LUCC data, a digital elevation model (DEM), and vector data. The LUCC data were obtained from the Global 30 m Land Cover Change Dataset (GLC_FCS30D) from 1995 to 2020 (https://data.casearth.cn/list?q=GLC_FCS30 (accessed on 3 July 2024)) [26]. DEM data were sourced from the Shuttle Radar Topography Mission (SRTM) (https://dwtkns.com/srtm30m/ (accessed on 15 October 2024)), also with a spatial resolution of 30 m. Vector administrative boundary data were acquired from OSM-Boundaries (https://osm-boundaries.com/map (accessed on 23 November 2024)). Using ArcGIS 10.8, the above datasets were mosaicked, cropped, and merged. LUCC types were classified into 8 categories (

Table 2. LUCC classification system and reclassification scheme in northern Oman.

Original Code	Classification System	Reclassification (Code)
10	Rainfed cropland	Arable land (1)
11	Herbaceous cover cropland
20	Irrigated cropland
61	Open deciduous broadleaved forest (0.15 < fc < 0.4)	Forest (2)
62	Closed deciduous broadleaved forest (fc > 0.4)
71	Open evergreen needle-leaved forest (0.15 < fc <0.4)
130	Grassland	Grassland (3)
120	Shrubland	Shrubland (4)
122	Deciduous shrubland
210	Water body	Water body (5)
190	Impervious surfaces	Impervious surfaces (6)
200	Bare areas	Sand (7)
201	Consolidated bare areas
202	Unconsolidated bare areas
150	Sparse vegetation (fc < 0.15)
152	Sparse shrubland (fc < 0.15)
153	Sparse herbaceous (fc < 0.15)
180	Wetland	Wetland (8)

) according to local conditions. All datasets were projected to the WGS_1984_UTM_Zone_39N coordinate system to ensure spatial consistency.

2.3. Methods

2.3.1. Vector Chain Tracing Method

LUCC transfer flow refers to the amount and direction of LUCC change between two time periods [27]. Continuous records of LUCC flows across multiple periods form LUCC transfer chains. Based on the cumulative contribution rate of LUCC change, representative chains are selected, their dynamic characteristics are summarized and classified, and corresponding spatial atlases are generated. A geoscience information atlas is a composite spatio-temporal analysis method that simultaneously depicts spatial structure and dynamic processes. It combines the attributes of both maps and genealogies: maps represent spatial location, while spectra (or chain diagrams) convey process-change information. Integrating the two enables the combined analysis of “space and process”.

LUCC transfer chain describes the sequential transformation of a LUCC type into other types at multiple time points. Using a forward-tracing approach based on the initial remote sensing image, the initial LUCC type is identified, and classification results from each period are sequentially overlaid to record change paths [28]. Change codes of the same type are merged to produce a unique chain code. To reduce storage requirements, years are encoded as letters in chronological order, and when a patch changes, a subscript marks the letter corresponding to the year, forming a simplified code. For example, if a patch’s LUCC code is 7 in the initial year, changes to 1 in 2000, reverts to 7 in 2005, and then changes to 4 in 2020, the complete chain code is 717774, while the simplified code is 7_A1_B7_E4. LUCC chains vary in length and structure, and are classified into one-way chains (Type A), reciprocal chains (Type B), and random chains (Type C). Type B is further subdivided into simple (B₁) and complex (B₂) reciprocal chains. Given the large variety but small quantity of random chains, this study focuses on one-way and reciprocal chains (

Table 3. Types and definitions of LUCC transfer chain types based on transformation pathways and sequence characteristics.

Species	One-Way Chains (Type A)	Reciprocating Chains (Type B)		Random Chains (Type C)
		Simple Reciprocating Chains (Type B1)	Complex Reciprocating Chains (Type B2)
Features	During the research period, the LUCC type changed once	During the research period, the LUCC type changed twice and reverted to its original type	During the research period, the LUCC type underwent multiple alternations between two categories	During the research period, the LUCC type underwent more than two consecutive changes across different categories
Coding method (Example)	Arable land converted to grassland code: 13	Arable land converted to grassland and then back to arable land code: 131	Alternating conversions between arable land and grassland code: 1313	Sand converted to arable land, then to impervious surfaces code: 716

).

To reveal the mechanisms underlying LUCC evolution, this study employs a vector tracing method. First, spatial atlas representation is combined with chain-type classification to enhance the structural identification of transfer paths based on the spatial distribution of change chains. Next, temporal tracking analysis is applied to typical chains to determine their initial LUCC type, change sequence, and terminal state, thereby enabling a complete spatio-temporal characterization of LUCC transfer chains. The construction of chain codes establishes a mapping between attribute dimensions and the spatial atlas, providing technical support for vector tracing and serving as a key step in spatial modeling of characteristic LUCC transfer chains.

In summary, the method integrates spatial atlas representation of LUCC characteristic chains with multiple analytical components, including the LUCC transfer matrix, LUCC transfer flow, LUCC transfer chain, and LUCC spatial atlas mapping. The analytical framework is illustrated (Figure 2). This process makes it possible to determine, from the transfer chain itself, the initial type, the sequence of conversions, and the final state at each stage. In this way, the increase in one LUCC type can be directly linked to the reduction in another, since the chain explicitly records how one category is transformed into another across different periods.

2.3.2. Spatial Analysis Methods

This study primarily applies hotspot analysis and kernel density estimation. In hotspot analysis, 30 m LUCC data are aggregated into 3 km grids to calculate the area proportion and relative change rate of each LUCC type, identifying hotspots where either the area or contribution exceeds 20% [29]. The Getis-Ord Gi* statistic is used to assess the spatial clustering significance of hot and cold spots. Kernel density estimation calculates the density of features within a defined neighborhood [30], generating smoothed surfaces that highlight areas of high and low density. Higher Kernel density estimation values indicate a greater spatial concentration of a given LUCC type. The formula is provided (

Table 4. Spatial analysis methods for LUCC change.

Method	Formula	Meaning of Letters
Hotspot Analysis	${G_i}^{*}={\displaystyle \frac{{\displaystyle \sum _{j=1}^n{w}_{ij}{x}_j-\overline{X}{\displaystyle \sum _{j=1}^n{w}_{ij}}}}{S\sqrt{\left[n{\displaystyle \sum _{j=1}^n{w}_{ij}^2}-({\displaystyle \sum _{j=1}^n{w}_{ij}{)}^2}\right]}}}$; $\overline{X}={\displaystyle \frac{{\displaystyle \sum _{j=1}^n{x}_j}}{n}}$; $S=\sqrt{{\displaystyle \frac{{\displaystyle \sum _{j=1}^n{x}_j}}{n}}}-{\left(\overline{X}\right)}^2$	Gi represents the spatial clustering of elements; xj denotes the attribute value of statistical grid j; wij is the spatial weight between grids i and j; $\overline{X}$ is the weighted mean; S is the weighted standard deviation; and n is the total number of grids.
kernel density estimation	$f(x)={\displaystyle \frac{1}{n{h}^2}}{\displaystyle \sum _{i=1}^{n}K\left(\frac{x-x_i}{h}\right)}$	f(x) denotes the kernel density estimation value; h is the search radius (smoothing parameter); n is the number of elements within the search radius of the kernel center; and x − xi is the distance from the estimation point x to the sample xi.

).

3. Results and Analysis

3.1. Spatio-Temporal Dynamics of LUCC Change

Northern Oman is predominantly covered by desert sand, while land suitable for agriculture, urban development, or with high ecological value occupies only a small proportion (Figure 3a). Arable land is concentrated in piedmont alluvial fan zones, impervious surfaces cluster around major coastal cities such as Muscat and Sohar, and shrubland and grassland are mainly restored along both flanks of the Al-Hajar Mountains and adjacent ecological buffer zones. LUCC hotspots during the study period were primarily located along the coasts of the Persian Gulf and Gulf of Oman—particularly in North Al-Batinah, Muscat, and northern Ad-Dakhiliyah—corresponding to densely populated piedmont fans, transportation corridors, and traditional settlement clusters that serve as core areas for agricultural and urban expansion (Figure 3b). In contrast, cold spots were identified on the western slopes of the Al-Hajar Mountains, at the margins of the Rub’ al Khali Desert, and in the northern interior of the Wahiba Sands. From 1995 to 2020, sand cover declined from 125,735.02 km² (99.03%) to 123,166.88 km² (97.01%). In contrast, arable land expanded from 644.10 km² to 1384.95 km² (a 2.15-fold increase), and impervious surfaces grew from 389.12 km² to 1250.02 km² (a 3.21-fold increase), reflecting concurrent expansion of oasis agriculture and urban construction (Figure 3c). Grassland increased from 92.11 km² to 336.10 km², and wetland area rose from 18.05 km² to 120.71 km², indicating localized ecological restoration. Overall, the region displayed a compound LUCC pattern characterized by oasis expansion, urbanization, and ecological recovery (Figure 3d). Sand was the dominant source land type, accounting for 89.38% of total transfer-out flows, primarily converting into arable land, impervious surfaces, and shrubland. The largest gains were recorded in arable land (942.30 km²), followed by impervious surfaces (758.35 km²) and shrubland (465.57 km²).

3.2. Characteristics of LUCC Transfer Chain-Spectrum

3.2.1. Identification of Key LUCC Transfer Chains

LUCC change in northern Oman was predominantly characterized by one-way transfer chains, which accounted for 69.76% of total change, followed by reciprocating chains (24.50%). Between 1995 and 2020, the primary transitions involved the conversion of sand into arable land, impervious surfaces, and shrubland (Figure 4). The most prominent pathways were sand to arable land (Type 71), sand to impervious surface (Type 76), and sand to shrubland (Type 74), contributing 941.17 km², 757.44 km², and 465.01 km², respectively, and together accounting for 70.11% of the total transition area. Secondary transitions included sand to grassland (Type 73, 292.48 km²) and sand to forest (Type 72, 181.43 km²). These five dominant pathways together represented 85.46% of total LUCC, serving as the key transfer chains in the region. By contrast, transitions among arable land, impervious surfaces, and ecological land types such as forest, grassland, and wetland were minimal (Type 16: 3.31%, Type 82: 2.89%, Type 12: 0.04%, Type 18: 0.01%). Spatially, Type 71 chains were concentrated in the coastal plains and piedmont zones of Al-Batinah North, Al-Batinah South, and Ash-Sharqiyah South, forming fan-shaped clusters at the coast-piedmont interface typical of oasis agricultural expansion. Type 76 chains occurred in urban centers such as Muscat and North Al-Batinah, as well as in point-pattern distributions across Ad-Dakhliyah, Ash-Sharqiyah South, Al-Wusta, and the western edge of South Al-Batinah, reflecting urban growth along transport hubs and township corridors. Type 74 chains were widely distributed along the northern and southern flanks of the Al-Hajar Mountains, particularly in Ad-Dakhliyah, Al-Batinah South, and the junction between Muscat and Ash-Sharqiyah South, forming core ecological restoration zones within mountain corridors and inter-oasis belts. Type 73 chains were relatively scattered, mainly in the western foothills of the Al-Hajar Mountains and central-southern Ash-Sharqiyah South, often located in transitional piedmont zones and low-elevation hills, indicating high restoration potential. Type 72 forest-restoration chains, though small in area, were concentrated at the western edge of Al-Batinah South and the highland junction with Ad-Dakhliyah, reflecting targeted ecological restoration. Overall, these key transfer chains formed an orderly spatial distribution across piedmont plains, urban corridors, and ecological boundaries. While sand remained the primary source type, its transition directions varied according to geomorphic setting and development context: arable land clustered in piedmont alluvial fans, impervious surfaces concentrated around coastal cities and transport nodes, and shrubland or grassland restoration occurred along mountain flanks and ecological buffers. Northern Oman thus exhibits a functional zonation pattern, serving as a critical region for both urban-agricultural expansion and ecological regulation.

3.2.2. Typical One-Way LUCC Transfer Chains

In typical one-way chains, the transformation from sand to arable land is densely distributed within oasis belts such as Al-Batinah North, Al-Batinah South, and Ash-Sharqiyah South, while the transition from sand to impervious surfaces is concentrated on the outskirts of urban areas such as Muscat (Figure 5). The expansion of urban and agricultural land is not mutually exclusive but reflects a coordinated spatial layout shaped by terrain and accessibility. Meanwhile, the localized distribution of chains converting sand into shrubland and grassland indicates ongoing ecological restoration in certain areas, contributing to the formation of a multifunctional LUCC system composed of “ecological buffer-agricultural core-urban node” zones. One-way chains are primarily characterized by conversions from sand to arable land, shrubland, grassland, impervious surfaces, and forest, with the most pronounced changes occurring between 2010 and 2020, particularly in northern coastal and inland transitional zones (Figure 5). Type 71 (sand to arable land) consistently ranked among the top transitions across five time periods. The 7_E1 chain accounted for 417.96 km² (11.09% of the total), indicating that agricultural development was especially prominent from 2015 to 2020, with large-scale expansion of arable land into sandy piedmont areas, progressively enlarging the oasis structure. Type 74 (sand to shrubland), mainly represented by chain 7_E4, covered 365.44 km² (9.70%), suggesting partial replacement of sand with drought-tolerant shrubs and initial recovery of ecological functions. Type 73 (sand to grassland), dominated by 7_E3, spanned 227.62 km², showing improved vegetation cover and clear restoration effects. Type 76 (sand to impervious surface) also exhibited high activity throughout all periods, with chains 7_D6, 7_E6, and 7_C6 contributing 5.45%, 4.64%, and 4.06% of total transitions, respectively, reflecting rapid urban growth along the northern margin of the Al-Hajar Mountains and around major towns. Type 72 (sand to forest) was notable only during 2015–2020, indicating considerable increases in vegetation cover in the Al-Hajar Mountains. The conversion of sand to arable land was driven by irrigation development in piedmont oases, while expansion of impervious surfaces was concentrated in transport-accessible coastal zones. In contrast, sand to shrubland transitions occurred mainly in ecological restoration zones along mountain corridors. Collectively, these one-way transfer chains depict the spatial evolution from a predominantly desert ecosystem toward a diversified landscape integrating oasis, agricultural and urban land, embodying the dual processes of ecological restoration and human development.

3.2.3. Typical Reciprocating LUCC Transfer Chains

Reciprocating chains are mainly characterized by mutual conversions between sand and arable land, grassland, or shrubland, and are predominantly distributed along the boundary zones of the Al-Hajar Mountains and adjacent provinces (Figure 6), particularly in the border areas of Al-Dhahira and Ad-Dakhliyah, as well as at the junction of Muscat, Ash-Sharqiyah North, and Ash-Sharqiyah South. Among these, Type 717 (sand to arable land then back sand) is the most representative, indicating consistently high land use intensity across all periods. In such cases, sand is initially reclaimed for cultivation but later reverts due to land degradation and the inability to sustain agricultural use. The typical chain 7_D1_E7 covered 76.95 km², reflecting excessive agricultural development in certain zones. Types 747 and 737 represent bidirectional transitions between sand and shrubland or grassland, with similar transfer volumes. Type 737, mainly involving chains 7_D3_E7 and 7_C3_D7, accounted for 1.23% and 0.76% of total transitions, respectively, while typical Type 747 chains: 7_B4_C7 and 7_C4_D7 contributed 37.89 km² and 29.67 km². After 2000, land quality in these areas showed partial improvement, but ecological conditions remained unstable. These chains are concentrated along the northern edge and eastern piedmont of the Al-Hajar Mountains, indicating that although shrubland and grassland once expanded, ecosystem stability still requires strengthening. Overall, reciprocating chains reflect localized land use instability, particularly the widespread occurrence of Type 717 chains, which represent pressure points caused by overdevelopment and insufficient ecological recovery. Most of these chains occur on the periphery of towns and at ecological boundaries rather than in core arable land or major urban expansion zones. This suggests that the broader land system retains substantial structural stability and resilience. Although some chains show signs of arable land degradation, they have not been extensively encroached upon by urbanization. Instead, under ecological regulation policies, certain degraded arable land has been partially restored. Nevertheless, the area of land degradation from sand remains significantly larger than the area of improvement, underscoring the continued need to balance agricultural development with ecological conservation.

4. Discussion

This study identified the key transfer chain-spectrum features of LUCC in northern Oman using the vector tracing method. Under the combined influence of natural stresses and human interventions, LUCC in arid regions generally exhibits cyclical fluctuations [31]. The results showed that the unidirectional transfer from sand to cropland and impervious surfaces was the dominant process, consistent with findings from central Asia regarding desertification control outcomes [32]. This indicates that concentrated transfers toward agricultural and urban land driven by human development are common in arid and semi-arid regions. Relative to the marginal areas of the Sahara in north Africa [33], the proportion of transfers from sandy land to grassland and woodland in this study was lower, highlighting the differences in regional water resources and climatic contexts [34]. The method effectively identified key nodes of urban expansion, ecological restoration, and land degradation, and revealed the synergistic relationships among urban expansion, oasis development, and ecological governance. Unlike the typical pathway of large-scale cropland encroachment observed in many arid regions [35], this region did not exhibit a widespread conversion of high-quality cropland into urban land. However, the substantial presence of “restoration-degradation” reciprocating chains in the spectral structure reflected cyclical conversions among cropland, ecological land, and sandy land, revealing the instability of ecosystem functioning. The findings are also consistent with previous studies in Oman: rapid urbanization and economic growth have been identified as the primary drivers of LUCC, while reduced precipitation and rising temperatures due to climate change have further accelerated the breakdown of traditional land cover patterns [36]. In recent years, urban land has increased significantly, while agricultural land and natural vegetation have declined due to human activities [37], and urban areas are expected to continue expanding in the future [38]. These results suggest that the stability of land conversion processes in Oman largely depends on sustained policy interventions and technological support.

The vector tracing method demonstrates unique advantages. Compared with static patch comparison analysis, by simultaneously considering the quantity, direction, chain pathways, and spatial expression of LUCC changes [39], it provides a more comprehensive representation of complexity and dynamics. Previous studies have confirmed its effectiveness in identifying the structures and interactions of LUCC transfers. For example, Xu et al. [40] used land transfer chains to reveal land class transitions in Jiangxi, China; Xie et al. [41] demonstrated that transfer quantity and direction determine LUCC characteristics in Guangxi’s rocky desertification areas of China; Wu et al. [42] investigated the spatiotemporal change flows of LUCC in central Ningxia, China; and research in the ecological functional zone of Lake Nasser in northern Egypt also confirmed these findings [43]. Compared with these studies, the present research further strengthened the comprehensive integration of temporal, structural, and spatial dimensions through a chain-spectrum collaborative framework, providing new perspectives for methodological applications. By integrating dynamic flow analysis, chain construction, and spectral representation, this approach enables full-sequence, multi-dimensional characterization of LUCC evolution while preserving complete transfer pathway information. Nevertheless, this method has certain limitations. First, it relies heavily on high-precision and long-term remote sensing data. In cases where classification accuracy is insufficient or data are missing, the construction of chain codes and identification of chain spectra may be affected. Second, the current analysis primarily focuses on natural attributes, with limited consideration of socio-economic factors. With the increasing availability of global-scale, annually resolved LUCC datasets and the growing integration of socio-economic and climate data, future research could advance multi-scale coupling and interdisciplinary analyses, providing opportunities to further improve coordinated studies of LUCC chain-spectral analysis. This will be a key direction for subsequent research.

5. Conclusions

This study focused on northern Oman, a representative area within the oasis–desert ecotone of Western Asia. Using multi-source remote sensing data, a vector-tracing method was developed to examine the chain-spectrum analysis of LUCC from 1995 to 2020. The main conclusions are as follows:

(1)

LUCC in northern Oman is dominated by desert, while the proportions of arable land, impervious surfaces, and key ecological land types remain relatively low. From 1995 to 2020, oasis areas expanded significantly, with arable land increasing from 0.51% to 1.09%. Urbanization also progressed rapidly, with impervious surfaces increasing from 0.31% to 0.98%. Desertification control yielded positive results, as sand cover declined from 99.03% to 97.01%. Spatially, oasis, impervious surfaces, and ecological land formed a functionally complementary layout: arable land was concentrated in piedmont irrigation zones, impervious surfaces were located near the coast and urban centers, and forest and grassland were mainly distributed along the Al-Hajar Mountains.

(2)

Human activities including oasis development, urban expansion, and ecological restoration were the dominant drivers of LUCC. Chain-spectrum analysis showed that 69.76% of land transitions were anthropogenically driven, particularly through unidirectional chains of sand converting to arable land (7_C1, 7_D1, 7_E1), impervious surfaces (7_C6, 7_D6, 7_E6), and shrubland (7_E4), highlighting the deep penetration of human development in LUCC evolution in arid regions.

(3)

Ecological constraints related to water scarcity and arid climate continued to play a major role. Limited water resources led to the formation of typical reciprocating chains in the oasis–desert interface (7_D1_E7, 7_A1_B7, 7_C1_D7), reflecting land degradation and fallback following farmland reclamation. Climatic effects were particularly evident in the arid vegetation zones along the Al-Hajar Mountain foothills, where sand erosion drove shrubland sand shrubland transitions (7_D3_E7, 7_C3_D7). These reciprocating chains accounted for 24.50% of LUCC, revealing the ecosystem’s sensitive feedback to climatic and hydrological stress.

(4)

The LUCC pattern in northern Oman effectively avoided spatial conflicts among oasis development, urban expansion, and ecological restoration. During the study period, the proportion of transfer chains among arable land, impervious surfaces, and ecological land types (forest, grassland, and wetland) remained low (Type 16: 3.31%, Type 82: 2.89%, Type 12: 0.04%, Type 18: 0.01%). This reflects the adaptive coupling of the human–land system and challenges the conventional narrative that urbanization in arid zones inevitably encroaches upon farmland. The case of northern Oman provides a transferable reference for optimizing land use structure and implementing differentiated zoning governance in ecologically fragile arid regions.