GIS-based assessment of land degradation vulnerability and environmental indices as ecological indicators in Western Agro-climatic zone of Tamil Nadu, India

Abstract

Land degradation is a critical global issue that affects soil fertility, food production, and biomass, driven by climate change and human activities. This study evaluates land degradation (LD) vulnerability across various land use and land cover (LULC) types in the Western Agroclimatic Zone of Theni district, Tamil Nadu, India. The GIS-based land degradation vulnerability index (LDVI) is calculated to identify site-specific degradation rates by combining four indicators (QI): Soil Quality Index (SQI), Climate Quality Index (CQI), Vegetation Quality Index (VQI), and Land Management Quality Index (MQI), derived from various geo-environmental and climatic variables. The LDVI map classifies the area into four categories: (i) non- affected zones (N), (ii) potential vegetative cover (P) with LDVI < 1. 22, (iii) fragile zones (F 1, F 2, and F 3) with LDVI values from 1. 23 to 1. 1.37, and (iv) critical zones (C 1 and C 2) with LDVI values between 1. 38 and 1. 53. A critical zone (C 3) with LDVI > 1.53. 53 indicates a severe land degradation risk, covering 3.68% of the total area. About 25. Approximately 38% of the region is affected by severe land degradation, primarily on pediplains, barrens, and fallows, resulting from soil erosion, salinity intrusion, nutrient loss, and inadequate land management. Meanwhile, 44. 38% falls within fragile categories, primarily on alluvial plains with red loamy soils. Conversely, 26. 3% of the area is classified as lower risk zones, such as forests, plantations, and irrigated lands, which benefit from soil fertility retention and effective management. The geographic correlation analysis reveals a strong positive relationship with VQI and MQI, with ‘r’ values of 0.835 and 0.831, respectively, indicating a risk to cultivable lands. Significantly, Pearson’s correlation confirms a strong positive relationship between LDVI and SQI and VQI, with coefficient values of 0.99 and 0.84, respectively, indicating that changes in soil properties and vegetation (NDVI) have a direct influence on land degradation across various regions. These findings provide a site-specific land degradation rate and its spatial relationship to quality indicators, emphasizing the importance of land-water-soil management for mitigation. Although the LDVI map reveals a spatial pattern of land degradation at a 30 m x 30 m pixel scale, limited by the resolution and temporal scale of the input datasets, future work could improve the LDVI by incorporating higher-resolution input data and field observations. This research directly supports UN-SDG 2 - Zero Hunger (sustainable agriculture and food security), and SDG 13 - Climate Action (climate resilience and adaptation).

Introduction

Land degradation represents a significant global challenge that impacts agricultural output and influences the socio-economic conditions of both local and regional communities. ^1–6. Land degradation can be understood as the transformation of productive land into desert-like conditions, where its physico-chemical and biological functions are reduced because of the combined effects of climate fluctuations and human-induced activities^7–11. UNEP (1983)¹² describes land degradation as the diminishing of ecosystem services and productivity. Such degradation impairs land, soil, and water resources, thereby lowering fertility and productivity, most evident in arid, semiarid, and dryland zones^13–16. UNCCD (2014)¹⁷ reported that nearly one-quarter of the Earth’s land area is vulnerable to desertification, driven by climate change and human pressures that significantly affect soil fertility, nutrient availability and agricultural productivity. Globally, the spread of land degradation has ultimately reduced soil fertility and food productivity in arid, semiarid, and subtropical areas^18–23. Land degradation is shaped by a complex interplay of natural conditions, such as soil composition, climate variability, and topography, alongside socio-economic pressures, including population growth, land tenure practices, and governance issues²⁴. Land degradation manifests in various forms, including erosion, vegetation loss, soil compaction, salinity accumulation, and declining soil fertility. These conditions spread across landscapes due to both natural factors, such as rainfall variability and climate change, and anthropogenic pressures, including encroachment, overgrazing, deforestation, burning, land abandonment, and poorly managed irrigation, resulting in lasting damage to land, soil, and water resources^25–28.

Land degradation has a significantly impact on diverse landscapes across Africa, Asia, South America’s central regions, Western Australia, North America, Europe, the Mediterranean, and the Sahara, leading to serious challenges such as increasing aridity and declining soil fertility caused by climatic factors^29–32. In India, about 20% of the landscapes, especially in arid and semiarid regions, fall under soil and land degradation impacts due to rainfall and climate variability³³. Climatic variability combined with poor land management contributes to land degradation, which is evident not only in semiarid areas but also in subtropical and semi-humid environments³⁴. SAC (2021)³⁵ has reported that the national-scale assessment of land degradation status indicates degradation by erosion (11.01%), vegetation degradation (9.15%), and wind erosion (5.46%), with a cumulative rate increasing to 1.87 Million Hectares between 2018 and 2019. In the Southern Indian landscapes, the degradation occurs in semiarid and sub-tropical regions, extensively in regions with black cotton soils (black clay and calcareous clay) and red loamy soils³⁶. Changes in landuse and land cover (LULC) over the agroclimatic zones have caused severe impacts on land-water-soil resources, contributing to soil erosion, surface runoff and salinity ingress, leading to land degradation, followed by desertification in larger areas, associated with rainfall variability and climate change^37–41.

Understanding land degradation is crucial for implementing effective mitigation measures and promoting sustainable land management practices. Remote sensing, combined with GIS, enables the assessment of land degradation, allowing researchers to analyze multiple contributing factors and their variations across both spatial and temporal scales. Satellite images provide spatio-temporal coverage of land-soil-water properties using Sentinel 2 MSI (10 m), IKONOS (1 m), Quick Bird (0.6 m), Landsat OLI (30 m), and IRS – LISS 3 (23.5 m), etc. The GIS techniques can be used to store, analyze, and retrieve geospatial data for land degradation zones^44–46. Many researchers have applied GIS-based methods to assess land degradation worldwide, including FAO/UNEP⁴⁷, MEDALUS^9,48,49, PESERA^50,51, LADA⁵², DISMED^53,54, IMDPA^55,56, AHP^56–58 and DesertWatch Extension^59,60. Machine learning (ML) and artificial intelligence (AI) methods, including Support Vector Machine (SVM), Gradient Boosting Machine (GBM), Generalized Linear Model (GLM), and Random Forest (RF), have increasingly been adopted to predict land degradation, employing empirical equations that incorporate multiple environmental and spatial variables⁶¹.

The MEDALUS model is widely used for assessing and mapping land degradation vulnerability in various parts of countries worldwide^37,40,62,63. The MEDALUS equation analyses quality indicators of soil, land, vegetation, and climate parameters. It has proved an effective tool for delineating land degradation zones in diverse landscapes^64,65. Dharumarajan et al. (2018)¹⁵ assessed about 30% of the land under the land degradation (LD) class in the Anandpur district of Andhra Pradesh. Dwivedi et al. (2024) have executed the MEDALUS model to assess the LD risk in the Satara and Sangli Districts of Maharashtra, India. Kumar et al. (2024)⁶⁷ have identified degradation zones along the Nandakini Watershed in the Himalayan region by incorporating the AHP technique and site-specific parameters like slope, terrain ruggedness, and soil organic carbon content. Rajbanshi and Das (2021)⁶⁸ monitored land sensitivity to desertification using the ESAI approach based on the MEDALUS model across India from 1992 to 2015, identifying that Rajasthan and Ladakh possess the highest mean ESAI values (1.5–1.7), where 87.61% and 83.83% of land, respectively, are critically degraded. Praveen et al. (2016)⁶⁹ conducted an assessment of land degradation vulnerability for conserving land quality in the north coastal areas of Tamil Nadu, analyzing the present and future possible land degradation under the purview of climate change impacts on the South Indian Coast. Kaliraj et al. (2022)⁷⁰ demarcated land degradation zones in Palakkad district (sub-tropical zone) and Virudhunagar district (semiarid zone) using the MEDALUS model, and these areas are facing a severe degradation process due to soil erosion, salinity ingress, and loss of soil nutrients.

The LULC features along the western agroclimatic zone of the Theni district are critically threatened by natural and anthropogenic factors, which are drastically inducing land degradation issues, due to soil erosion, intensive rainfall after prolonged summer, improper landuse management practices, and are causing adverse impacts on soil infertility, removal of nutrients, imbalanced soil-water, and reduced agriculture productivity. Therefore, it is vital to assess land degradation through scientific approaches and multi-parameter analysis to ensure sustainable land-soil-water resource management, thereby improving the livelihoods of local communities. In response to these challenges, this study aims to develop a Land Degradation Vulnerability Index (LDVI) using multiparametric analysis within a GIS framework, integrating quality indicators of soil, climate, vegetation, and land management. It further aims to delineate and classify the spatial variability of land degradation into non-affected, potential, fragile, and critical zones and to examine how the contributing quality indicators influence the LDVI patterns across the study area.

Study area

The study area covers the Western Agroclimatic Zone of India, located along the eastern slope of the Western Ghats in the Theni District of Tamil Nadu, India (Fig. 1). The region covers 2,869 km², with its geographical extent lying between the latitudes of 9°32’00’’N and 10°15’00’’N, and the longitudes of 77°10’00’’E and 77°40’00’’E. The Theni district comprises five taluks, six municipalities, and 130 villages, with a population of approximately 1.245 million. The major drainage system, namely the Vaigai and Suruli rivers, flows from the hills of the Western Ghats towards the northeastern plains, exhibiting a dendritic drainage pattern⁷¹. The landscapes with the relief range of 160–1,400 m above MSL, wherein the Ghats’ upland hill ranges (ridges and valleys) cover the west, south, and southeast parts, and alluvial plains spread over the middle and northeastern parts with vast settlements and agricultural lands. Geologically, the major portions underlie the charnockite with irregular hillocks, while alluvial plains are associated with hornblende-biotite gneiss and sillimanite-garnet-biotite gneiss, interlined by minor granite patches. Quaternary sediment deposits, mostly alluvium, are found along the central and northeastern regions. Charnockite covers around 70% of the terrain, followed by migmatite and Khondalite, and there is substantial stone mining and quarrying activity. The district’s soil composition comprises red loamy, red sandy, red gravelly, black clay, and brown soils. Predominantly, red gravelly and red loamy soils occur, exhibiting moderate to high permeability and sediment layers ranging from medium to coarse grains. Low permeability is found in isolated regions of black clay and brown soils, which account for only 1% of the total area. Alluvial soils are primarily found in floodplains along rivers. The district’s land use and land cover (LULC) include irrigated cultivable lands, dry farmland, plantations, barren lands, steep terrains, shrublands, urban and rural communities, and forest cover. Cultivation occupies over 40% of the land, while forest and barren regions account for 34% and 26%, respectively. Groundwater sources vary depending on the hydrogeological settings, and groundwater can be found at depths ranging from 500 to 750 m in difficult rocky terrains and from 100 to 300 m in sedimentary formations. The groundwater table around bodies of water and riverbeds ranges between 2 and 20 m deep. The northeastern and northwest regions have the greatest groundwater potential due to their extensive agricultural activities. Situated within the Western Agro-climatic Zone, the Theni District has a temperate climate with temperatures ranging from 19 °C to 39 °C. Annual rainfall, contributed by both southwest and northeast monsoons, varies from 720 to 860 mm. Although agriculture supports the majority of the population, certain regions experience significant soil and land degradation due to both environmental and human-induced factors. Recent assessments by the ICAR–Indian Institute of Soil and Water Conservation (IISWC)⁷² highlight that Theni District faces significant land degradation challenges, with approximately 185.1 thousand hectares (64.55% of the total geographical area of 286.8 thousand hectares) requiring priority treatment for soil erosion management. The district falls under Severity of Risk Category B, indicating a critical extent (50,000–100,000 ha) that warrants immediate conservation intervention. The spatial analysis of erosion risk reveals that about 2.2 thousand hectares fall under Priority Class 1 (extremely high priority; erosion-tolerance difference > 35 t ha⁻¹ yr⁻¹), 27.6 thousand hectares under Priority Class 2 (25–35 t ha⁻¹ yr⁻¹), and 36.9 thousand hectares under Priority Class 3 (15–25 t ha⁻¹ yr⁻¹), together accounting for 22.2% of the district’s total area that requires urgent conservation measures.

Fig. 1.

The location map of the study area shows NDVI sample sites and major settlements overlaid on the Landsat 8—OLI image (True Colour Composite of VNIR bands). The map was generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Materials and methods

The GIS-based LDVI model is designed with a two-tier framework to execute multiple geo-environmental parameters and compute the Land Degradation Vulnerability Index (LDVI) equation at the pixel level, thereby demarcating land degradation vulnerable zones within the study area. Initially, four quality indices are derived: Soil Quality Index (SQI), Climate Quality Index (CQI), Vegetation Quality Index (VQI), and Management Quality Index (MQI). Each QI is computed using corresponding empirical equations (Eq. 2 – Eq. 5) from various geo-environmental and climatic parameters. The data sources and their derived Parameters used for LDVI assessment are illustrated in Table 1. In this analysis, each parameter has been represented in the thematic maps with its corresponding legend (attributes) to illustrate its spatial characteristics using ArcGIS v.10.6 software tools, based on the author’s digitization processes. Subsequently, the LDVI is calculated to demarcate the areas potentially vulnerable to land degradation on a site-specific scale, using the integrated GIS and remote sensing techniques^{21,37,38,45,53,73}. Figure 2 illustrates the methodology workflow used for land degradation mapping. The LDVI is calculated by summing the weighted quality index (QI) raster layer using the Raster Calculator tool in ArcGIS v. 10.6 software. The resulting output map indicates site-specific land degradation status and is primarily based on the quality indices and their respective input parameters.

Table 1.

Data sources and derived parameters used for LDVI assessment.

Index	Parameter	Source/dataset	Agency/portal	Spatial resolution/type
Soil Quality Index (SQI)	Soil depth	District Soil Map and field attribute data	TNAD and NBSS & LUP	Vector layer (author’s digitized, polygon)
	Soil texture	District Soil Map	TNAD and NBSS & LUP	Vector layer (author’s digitized, polygon)
	Soil drainage	District Soil Map and field attribute data	TNAD and NBSS & LUP	Vector layer (author’s digitized, polygon)
	Parent materials	District Resource Map (Geology layer)	Geological Survey of India (GSI)	Vector layer (author’s digitized, polygon)
	Slope gradient	SRTM DEM (30 m)—Open Data Product	USGS Earth Explorer (https://earthexplorer.usgs.gov)	Raster (30 m)—ArcGIS v.10.6 —Spatial Analyst tool
Management Quality Index (MQI)	Landuse/Land cover (LULC)	Landsat 8—OLI/TIRS image (Open Data Product)	USGS Earth Explorer (https://earthexplorer.usgs.gov)	LULC classified vector (30 m)—ArcGIS v.10.6 —Supervised Classification tool
	Land capability (LC)	Derived from LULC, soil AND slope integrated layers	TNAD field attribute data	Vector layer (30 m)
Vegetation Quality Index (VQI)	NDVI	Derived from Landsat 8—OLI (R and NIR bands)	USGS Earth Explorer (https://earthexplorer.usgs.gov)	Raster (30 m)—ArcGIS v.10.6 —Raster Calculator tool
	Fire risk	Derived from Landsat 8—OLI VNIR bands	USGS Earth Explorer (https://earthexplorer.usgs.gov)	Raster (30 m)—ArcGIS v.10.6 —Raster Calculator tool
	Drought resistance	Derived from NDVI and LULC integrated layers	USGS Earth Explorer (https://earthexplorer.usgs.gov)	Raster (30 m)—ArcGIS v.10.6 —Raster Calculator tool
	Erosion resistance	Derived from NDVI, slope, and soil texture integrated layers	Landsat 8—OLI, SRTM DEM, and TNAD soil map	Raster (30 m)—ArcGIS v.10.6 —Raster Calculator tool
Climate Quality Index (CQI)	Aridity index	Global Aridity Index (Global-Aridity_ET0 data)—(Open Data Product)	CGIAR–CSI GeoPortal (https://cgiarcsi.community)	Raster (30 m)
	Evapotranspiration	Global Reference Evapotranspiration (Global_ET0)—(Open Data Product)	CGIAR–CSI GeoPortal	Raster (30 m)
	Rainfall	IMD Rainfall data	India Meteorological Department ((https://dsp.imdpune.gov.in/index.php)	Raster (30 m)—ArcGIS v.10.6 —IDW tool

Fig. 2.

The flowchart shows the systematic workflow of the LDVI model.

Data collection and thematic layers preparation

The LDVI parameters and quality indicators (QI) were extracted from multiple geo-environmental and climatic sources, including Survey of India topographic maps (scale 1:25,000), Geological Survey of India (GSI) published geology and geomorphology maps, Landsat 8 OLI images, and SRTM DEM (30 m resolution). Soil maps and in-situ measured data (soil depth, texture, and drainage) were derived from the Tamil Nadu Agriculture Department (TNAD) and National Bureau of Soil Survey and Land Use Planning (NBSS&LUP) published data sources. Furthermore, data on cropland areas, land management practices, population, water storage, and other relevant factors were derived from district statistical reports. The decadal rainfall data (2000–2023) were collected from the IMD data portal (https://dsp.imdpune.gov.in/index.php). Landsat images (path 143, row 53) were retrieved from the USGS Earth Explorer web portal (https://earthexplorer.usgs.gov). The derived thematic layers included Normalized Difference Vegetation Index (NDVI), LULC, land capability, drought tolerance, fire-risk zones, and erosion resistance. The base map layers, such as administrative boundaries and drainage, were created from topographical maps produced by the Survey of India. IMD rainfall data was analyzed using the GIS-based Inverse Distance Weighting (IDW) tool to estimate average rainfall. The soil map was digitized using the published data sources of the Tamil Nadu Agriculture Department and the NBSS & LUP, supplemented with attributes such as soil depth, texture, and drainage data. The aridity and potential evapotranspiration (PET) indices were derived from the high-resolution (30 m) Global Aridity Index (Global-Aridity_ET0) and Global Reference Evapotranspiration (Global_ET0) datasets available through the CGIAR-CSI GeoPortal. The thematic layers were derived from various data sources, including soil, rainfall, LULC, vegetation, land use practices, and climatic factors. All thematic layers were converted to the UTM-WGS 84 projection and resampled to a uniform pixel size (30 m x 30 m) with a common spatial scale (1:6,500) using the resampling analysis in GIS software. The rasterized layers of SQI, CQI, VQI, and MQI indicators are converted to a common pixel size and scale and then input into the empirical equation to calculate the LDVI using the ArcGIS v.10.6 - Raster Calculator Tool.

Computing quality indices (QI)

The Land Degradation Vulnerability Index (LDVI) is an empirical equation that integrates the geometric mean of four quality indices (QI), including the Soil Quality Index (SQI), Climate Quality Index (CQI), Vegetation Quality Index (VQI), and Management Quality Index (MQI). Each Quality Index is derived by integrating multiple spatial parameters that collectively characterize the respective environmental components. These parameters are standardized on a vulnerability scale ranging from 1 to 2, wherein a value of 1 indicates low vulnerability (favorable conditions) and 2 signifies high vulnerability (unfavorable conditions). Parameter weights, ranging from 1 to 5, are assigned based on expert judgment and established literature to reflect their relative importance within each index. The weighted parameters are subsequently reclassified and processed in the GIS environment through raster-based overlay analysis to generate a thematic map for each Quality Index at a 30 m spatial resolution. The LDVI map is then computed as the geometric mean of these four spatial indices, where higher LDVI values correspond to greater susceptibility to land degradation, while lower values denote areas with minimal or no degradation risk^{2,20,21,32,74,75}.

The Quality Index (QI) score is calculated from the weighted values assigned to the feature class parameters (P1, P2, P3…Px) associated with the four quality indicators (SQI, CQI, VQI, and MQI) using the following equation (Eq. 1)⁷⁶,

1

Where QI is referred to as the Quality Index score, P_1, P_2, P_3…P_n is the parameters (thematic layers) of Quality Indicators; s is the index score; and n is the number of parameters used to calculate QIs. Secondly, the Quality Index (QI) of each indicator is calculated using the appropriate empirical Eq⁷⁷.

SQI represents the influence of pedological and topographic conditions on degradation processes. The parameters considered include soil depth, texture, drainage, parent material, and slope. Soil characteristics were extracted from the NBSS & LUP soil maps and reclassified based on their influence on infiltration and soil retention capacity. Parent material was derived from the GSI geological map, while slope was computed from the SRTM DEM using the slope function in GIS software. These layers were integrated to produce the SQI map, where shallow soils, poor drainage, hard rock formations, and steep slopes indicate higher vulnerability.

2

CQI reflects the influence of climatic constraints on land degradation and was generated using rainfall, aridity, and potential evapotranspiration (PET) parameters. Rainfall data (2000–2023) from IMD stations were spatially interpolated using the Inverse Distance Weighting (IDW) method to produce a continuous rainfall surface. The Global Aridity Index (Global-Aridity_ET0) and Global Reference Evapotranspiration (Global_ET0) datasets, obtained from the CGIAR–CSI GeoPortal, represent long-term climatological averages (1970–2000) and were downloaded in GeoTIFF raster format (30 m spatial resolution). The areas characterised by low rainfall, high aridity, and elevated PET values were categorised as more vulnerable to degradation. In contrast, regions with adequate rainfall and low aridity were considered less vulnerable to drought.

3

The Vegetation Quality Index (VQI) evaluates the condition, density, and resilience of vegetation, as well as its ability to protect the land surface from erosion and degradation. It was computed using four spatial parameters: Normalized Difference Vegetation Index (NDVI), drought resistance, fire resistance, and erosion resistance, all derived from Landsat 8 OLI imagery (30 m resolution). The NDVI was calculated using the standard band ratio from Bands 5 (NIR) and 4 (Red). Drought resistance was assessed by analyzing seasonal NDVI variations and identifying areas with persistent vegetation cover under low-rainfall conditions, indicating higher resilience. Fire resistance was derived from the Tasselled Cap Brightness and thermal infrared (Band 10) layers, where regions showing higher surface temperatures and low vegetation moisture were assigned higher vulnerability. Erosion resistance was computed through a GIS-based integration of NDVI, slope, and soil texture layers to represent the combined influence of vegetation density and soil stability on erosion susceptibility.

4

MQI captures the influence of human activities and land management practices on degradation. It integrates Land Use/Land Cover (LULC) and land capability parameters. LULC was mapped using supervised classification (Mahalanobis method) of Landsat 8 imagery, while land capability was derived from the integration of soil and slope data. Poorly managed croplands, fallows, and barren areas correspond to higher vulnerability, whereas forests and plantations represent conditions of low vulnerability.

5

The quality indicator’s QI score (1–2) is classified into three classes to represent the risk level of land degradation (LD). Table 2 presents the vulnerability class and rate of the LDVI model, along with its descriptions based on the Salvati and Sabbi (2011) standard classification systems.

Table 2.

LDVI vulnerability classes and ratings based on (Salvati & Sabbi, 2011).

Class	Subclass	LDVI ratings	Characteristics
Non-affected zone	N	< 1.17	Non-sensitive or very low-sensitive regions refer to lands that remain largely unthreatened or only slightly affected by degradation processes.
Potential vegetative cover	P	1.17–1.22	Lands designated as low-sensitive are potentially vulnerable to degradation when exposed to extreme climatic events or severe changes in land management practices.
Fragile zone	F1	1.23–1.26	Medium-sensitive lands lie at the brink of degradation, where perturbations in the fragile interaction between environmental factors and human activity may result in swift land deterioration.
	F2	1.27–132
	F3	1.33–1.37
Critical zone	C1	1.38–1.41	Regions classified as high or very highly sensitive to land degradation are substantially degraded, with marked declines in land productivity and ecosystem services.
	C2	1.42–1.53
	C3	> 1.53

GIS-based LDVI calculation

The GIS-based LDVI model calculates the Land Degradation Vulnerability Index (LDVI) by summing the quality index (QI) values of four quality indicators, including Soil Quality Index (SQI), Climate Quality Index (CQI), Vegetation Quality Index (VQI), and Management Quality Index (MQI). To compute the LDVI at the pixel level, the quality indicators were derived from multiple geo-environmental parameters, including soil, rainfall, LULC, vegetation, land use practices, and climatic factors (Kosmas et al., 1999). Despite the variability in scale and spatial resolution of the data sources, all thematic layers were converted into the UTM-WGS 84 projection and coordinate system and resampled into a uniform pixel size to ensure a consistent scale using the GIS Spatial Analysis Tools. The LDVI equation is computed the sum of QI values of SQI, CQI, VQI, and MQI indicators, which represent the rate of vulnerability to land degradation at a pixel-by-pixel scale. The LDVI is computed using the empirical equation (Eq. 6) in the GIS software environment. To execute this analysis, the thematic layers were assigned weights of 1–5 based on their sensitivity to land degradation and converted into a raster layer with a uniform pixel size (30*30 m) with a scale of 1:6.5 km. In which the weighted QI layers are summed to obtain the LDVI value using the GIS - Raster Calculator tool, employing the equation (Eq. 6), and is expressed as,

6

The LDVI result map is divided into three classifications to indicate the potential for land degradation: (i) non-affected zone (indicating potential vegetative cover), (ii) fragile zone (subdivided into F1, F2, and F3), and (iii) critical zone (subdivided into C1, C2, and C3). Each class of the LDVI map is quantitatively assessed. The calculated LDVI value indicates the cumulative rate of LD vulnerability at each pixel; thereby, a higher value indicates that the land with severe vulnerability to LD, while a lower value indicates the non-affected land, and the least vulnerability to land degradation.

Spatial correlation analysis of LDVI and quality indicators

The association between LDVI and QI layers was statistically evaluated through GIS - geostatistical tools, with Pearson’s correlation coefficient calculated to analyze their attribute data. Scatterplots from the spatial correlation analysis demonstrate the linear relationship between LDVI and QI, utilizing the least-squares regression method. The spatial correlation of contributing parameters is computed using the following Eq. 8¹ (Eq. 7), and it is represented as,

7

In this analysis, y is the dependent variable, and x is the independent variable (LDVI). The slope of the line is represented by m, and the y-intercept by c. Pearson’s correlation coefficients (r) were calculated to evaluate the magnitude and direction of linear associations, and R² values were used to determine how much of the variability in each dependent index could be attributed to the LDVI. This combined statistical and visual approach provided insights into the spatial correlations influencing land degradation vulnerability. Each thematic layer and its attributes were assigned weights and a quality index (QI) score, based on their degree of potential for degradation as presented in Table 3.

Table 3.

The LDVI parameter-wise assigned weights and quality index (QI) score⁷⁷.

Indicators	Parameters	Description	Assigned weights (threshold)	Index score (QIs)
Soil Quality Index (SQI)	Soil depth	Very deep	1	1
		Deep	2	1.1
		Moderately deep	3	1.5
		Moderately shallow	4	1.7
		Shallow	5	2
	Soil texture	Clayey	1	1
		Loamy	2	1.1
		Calcareous clay	3	1.5
		Gravelly clay	4	1.7
		Gravelly loamy	5	2
	Drainage	High permeability	1	1
		Moderate	2	1.2
		Lower	3	1.4
		Poor	4	2
	Parent materials (geological feature)	Alluvium	1	1
		Laterite	2	1.1
		Pink Granite	3	1.7
		Charnockite	4	1.8
		Granite gneiss	5	2
	Slope (degree)	< 6	1	1
		6–12	2	1.1
		12–24	2	1.2
		25–30	3	1.4
		30–35	4	1.6
		> 35	5	2
Climate Quality Index (CQI)	Rainfall (mm)	280–650	1	1
		> 650	2	1.2
	Aridity	< 0.50	1	1
		0.50–1.15	3	1.4
		> 1.15	5	2
	Evapotranspiration (mm)	< 1500	1	1
		1500–2000	3	1.4
		> 2000	5	2
Vegetation Quality Index (VQI)	Drought resistance	Very High	1	1
		High	2	1.4
		Moderate	3	1.5
		Low	4	1.6
		Very Low	5	2
	Fire resistance	Low	1	1
		Moderate	3	1.4
		High	5	2
	Erosion resistance	Very high	1	1
		High	2	1.4
		Moderate	3	1.6
		Low	4	1.7
		Poor	5	2
	NDVI	> 0.72	1	1
		0.51–0.72	2	1.2
		0.27–0.50	3	1.4
		0.14–0.26	3	1.6
		0.11–013	4	1.8
		< 0.1	5	2
Management Quality Index (MQI)	Land Use	Forest, Settlements, Waterbodies	1	1
		Agriculture irrigated,	2	1.6
		Agriculture un-irrigated, land with scrub	3	1.7
		Fallows, Sandy deposits	4	1.8
		Barren land, rocky outcrops, stony waste	5	2
	Land capability	Class 1	1	1
		Class 2	2	1.1
		Class 3	2	1.2
		Class 4	3	1.4
		Class 5	4	1.6
		Class 7, Class 8	5	2

Results and discussion

The LDVI map has been generated by summing the quality indicators, including SQI, CQI, VQI, and MQI, which were calculated using the GIS-based LDVI model, with a detailed elaboration on the characteristics of these parameters.

Soil quality index (SQI)

Soil quality is influenced by factors such as soil depth, texture, and productivity, and is crucial for supplying essential nutrients that support plant growth^77,80,81. Soil texture affects key properties such as erodibility, water retention, and stability of aggregates. Loamy soils typically have low water retention and nutrient content, while clay soils retain water and nutrients more effectively. Figure 3(a) depicts the dominant soil types, including red loamy soils, clay (black cotton), and calcareous clay, with smaller patches of gravelly clay, gravelly loam, and stony gravels. Figure 3(b) shows the parent materials like alluvium, laterite, granite, charnockite, and granite gneiss. Alluvial loams were deposited by rivers and streams, whereas clayey soils developed from the weathering of volcanic rocks such as gneisses and schists, combined with organic and inorganic matter. Figure 3(c) displays the regional distribution of soil depth across different soil types, divided into five categories: intense (> 3 m), deep (2–3 m), moderately deep (1–2 m), relatively shallow (1–2 m), and shallow (< 1 m). Soil depths are often greater in alluvial plains along riverbeds and mid-land pediplains, but shallower depths are found in uplands and hilly terrains. The slope of the ground impacts the topography, impacting soil thickness, permeability, runoff rates, and erosion by controlling the lateral transit of soil particles and solutes. Figure 3(d) classifies slope gradients as very steep (> 30 degrees) in valleys and hilly terrains, sharp (24–30 degrees) in upland forest areas, moderately sloped (12–24 degrees) in the foothills, and softly sloped (< 12 degrees) in the broader alluvial plains. These differences significantly impact the lateral transport of soil particles and solutes through surface runoff and overland flow. Soil drainage refers to the soil’s ability to store water, which affects soil-water transfer, infiltration, and accessibility for plant roots. Figure 3(e) depicts the spatial characteristics of soil drainage capacity based on different soil types, suggesting that alluvial plains possess higher soil-water drainage capacity. In contrast, shallow soils in uplands, stony waste, foothills, and barren lands tend to have limited drainage capacity. Lower soil-water content in these places increases the likelihood, particularly in alluvial plains and valley depressions³⁸.

Fig. 3.

Soil Quality Index (SQI) parameters: (a) soil texture, (b) parent materials (geological features), (c) soil depth, (d) slope gradient, and (e) soil drainage capacity. The maps were generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

The Soil Quality Index (SQI) reflects the spatial variability of soil quality and its resistance to localized land degradation⁸². Figure 4 categorizes soils into high/good quality (SQI < 1.13), moderate quality (1.13–1.45), and low quality (SQI > 1.45). Low-quality soils, covering 29.03% (833.01 km²), are predominantly found in uplands, rocky outcrops, stony waste, and barren lands, where shallow soils and poor water drainage hinder productivity. Moderate-quality soils occupy 65.64% of the district, primarily in mid-alluvial plains, characterized by red loamy, clay, and calcareous clay soils that are suitable for cultivation. High-quality soils, representing only 1.39% of the area, are localized near river alluvium and are rich in humus and organic content. Examination of the SQI indicators reveals that soil quality is comparatively lower in specific regions, particularly in upland and fallow areas, due to erosion and nutrient scarcity. These factors contribute to poor crop productivity in unirrigated croplands, shrublands, and barren lands.

Fig. 4.

The spatial characteristics of the Soil Quality Index (SQI). The map was generated in ArcMap v.10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Climate quality index (CQI)

The Climate Quality Index (CQI) evaluates the influence of rainfall, potential evapotranspiration (PET), and the aridity index (AI) on land degradation processes. Figure 5(a) and (b) illustrate the key parameters contributing to the CQI. Decadal average rainfall data from the India Meteorological Department (IMD) were used, covering the period from 2000 to 2023. Figure 5(a) illustrates the spatial variation in rainfall, with locations in the uplands and foothills of the southern and northwestern regions receiving over 650 mm of rainfall. In the mid-plains, a moderate rainfall rate of 280 to 650 mm is seen, progressively decreasing toward the northeastern portion while increasing in the northern and southern parts. Figure 5(b) illustrates the district’s Aridity Index (AI), derived using the rainfall ratio to the evapotranspiration rate. The AI is low in the alluvial plains of the middle and northeastern regions (< 0.5), indicating dry-humid conditions due to water stress and high PET. Conversely, irrigated lands and plantations display a moderate AI score (0.5 to 1.15), indicating sub-humid conditions with sufficient water for plant growth and development. The AI value reaches 1.15 in the upland forest cover and foothills in the western and southern regions, indicating heightened aridity. This increasing aridity enhances soil water stress, significantly affecting plant growth by altering evaporation and soil moisture processes (Kosmas et al., 1999), particularly damaging plants on the alluvial plains. The spatial variability of potential evapotranspiration (PET) is shown in Fig. 5(c). Elevated PET rates exceeding 2000 mm are recorded across the alluvial plains and calcareous clay soils. In contrast, moderate values between 1,500 and 2,000 mm are characteristic of foothill regions, plantations, and irrigated agricultural lands. Lower PET rates (< 1500 mm) are seen in the forested areas and plantations of the western and southern hilly regions. The variety in PET influences soil moisture, which in turn impacts plant growth and land productivity throughout the district.

Fig. 5.

Climate Quality Index (CQI) parameters: (a) Decadal average rainfall, (b) Aridity Index, and (c) Evapotranspiration. The maps were generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Figure 6 presents the Climate Quality Index (CQI) map, highlighting the spatial characteristics that indicate climatic pressures affecting the district’s land, soil, and water. A higher CQI (greater than 1.50) signifies lower climate quality and covers approximately 37.45% of the total area. This region primarily includes drylands and alluvial plains in the district’s central, eastern, and northeastern parts, where critical climatic variables are associated with lower soil-water content and a higher evapotranspiration rate. A moderate CQI (ranging from 1.25 to 1.50) accounts for 25.48% of the area, predominantly found in extensive irrigated lands characterized by loamy and black clay soils, especially in riverine alluvial plains. Approximately 33.14% of the study area falls within the high climate quality zone, with CQI values ranging from 1.00 to 1.25. This zone encompasses upland forests, foothills, riverbed lands, and plantations. These landscapes benefit from adequate rainfall, exhibit a low aridity index, and maintain a balanced soil-water content throughout all seasons. Areas with CQI values above 1.50 exhibit low climatic quality and a higher degradation risk, driven by elevated PET and high aridity. In contrast, CQI values below 1.25 correspond to favorable climatic conditions, characterized by adequate rainfall and lower water stress, which results in reduced susceptibility to degradation.

Fig. 6.

Climate Quality Index (CQI) map of the study area. The map was generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Vegetation quality index (VQI)

The spatial characteristics of VQI indicate the vegetative cover acting as a resilience factor against land degradation by controlling drought, erosion, fire risk, and moisture (soil-water content), retaining biomass across diverse climatic situations. Figure 7(a) illustrates the fire danger resistance and recovery potential of various land use and land cover (LULC) types. Fire risk is categorized into three classes according to vegetation, highlighting differences in plant sensitivity to fire. These variances affect soil quality, such as water repellency and nutrient loss, ultimately influencing runoff and erosion. The central and northeastern plains, comprising unirrigated croplands, shrublands, and barren areas, exhibit higher sensitivity to fire danger due to increased land surface temperature (LST) and insufficient rainfall, which results in decreased soil moisture and biomass moisture during the summer. Moreover, the upland forests and plantations exhibit exceptional resistance and recovery abilities against fire risk, which is attributed to frequent rainfall and ideal topographic wetness conditions. Moderate fire risk resistance has been seen in foothills, valley fills, grasslands, and associated LULC characteristics, which are predominantly driven by monsoonal rainfall.

Fig. 7.

Vegetation Quality Index (VQI) parameters: (a) Resistance capacity to fire risk, (b) Erosion resistance, (c) Drought resistance, and (d) NDVI of vegetation health. The maps were generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Figure 7(b) emphasizes the erosion resistance capacity of LULC features against hazards from water, runoff, and aeolian processes. Erosion resistance analyzes a vegetation’s ability to protect soil layers from erosion by reducing runoff, which is crucial for preventing land degradation. Throughout the area, soil fertility closely linked to forest cover, plantations, and irrigated croplands demonstrates increased resilience to erosion, effectively maintaining soil layers in all seasons. Conversely, pockets of croplands, fallows, and shrubland within slopes and alluvial plains demonstrate significant resistance to erosion. However, soils in alluvial plains, fallows, barren land, unirrigated croplands, and foothills are characterized by lesser erosion resistance, indicating a greater danger of erosion. This triggers the removal of topsoil, followed by nutrient loss, soil infertility, and related effects. Figure 7(c) displays the drought resistance capacity of LULC features, which is crucial for measuring land and soil degradation. Areas exhibiting high drought resilience occur in upland forests, plantations, and irrigated croplands, where sufficient rainfall maintains soil moisture, humus content, and humidity. Moderate drought resistance is observed in major alluvial plains, including unirrigated croplands and fallow lands, influenced by seasonal variations in land surface temperature (LST), soil moisture, and rainfall. The areas dominated by bare land, fallows, and shrublands exhibit lower drought resilience, which substantially reduces biomass productivity due to the combined effects of erosion and soil salinity.

Figure 7(d) illustrates the spatial distribution of the Normalized Difference Vegetation Index (NDVI), offering insights into variations in plant cover and vegetation health. The NDVI layer tracks vegetation greenness and health over time, enabling the identification of areas experiencing vegetative stress. NDVI values range from − 1 to 1, reflecting the level of photosynthetic activity and chlorophyll concentration in the vegetation. An NDVI value greater than 0.51 indicates healthy vegetation zones, including forest cover, plantations, and irrigated croplands. Values between 0.27 and 0.5 reflect moderately stressed vegetation, particularly in alluvial plains. NDVI values between 0.10 and 0.26 reflect moisture-stressed plants, including fallows, barren land, shrublands, and other degraded landforms. Values below 0.10 indicate the presence of non-vegetative features, such as built-up regions, water bodies, and rocky outcrops, throughout the research area.

Figure 8 displays the spatial characteristics of the Vegetation Quality Index (VQI), highlighting areas affected by vegetation degradation resulting from variables such as drought, erosion, fire danger, and other degradation processes. The areas with a VQI value of less than 1.15 indicate good vegetation quality, encompassing 30.08% of the region, primarily in upland forests and plantations. These places are characterized by dense vegetation, comprising both green and deciduous herbaceous layers, and they show resilience against drought, fire, and erosion throughout the seasons. A moderate VQI value of 1.13 to 1.38 is reported in 29.14% of the region, mainly in foothills, plantations, and irrigated croplands along riverbeds and waterways. Meanwhile, approximately 36.85% of the district falls into the low vegetation quality group (VQI greater than 1.38), mostly comprised of unirrigated croplands and barren lands in the alluvial plains. The data reveal that major areas of the district have seen severe vegetation degradation, resulting in moisture-stressed plants due to decreasing rainfall and poor soil-water retention conditions.

Fig. 8.

Vegetation Quality Index (VQI) map of the study area. The map was generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Land use management quality index (MQI)

The Land Use Management Quality Index (MQI) exhibits the influence of site-specific LULC features and their management practices on land degradation in conjunction with anthropogenic factors. The MQI values vary across different LULC classes, reflecting potential degradation risks stemming from specific land management practices. Based on the National Bureau of Soil Survey and Land Use Planning (NBSS-LUP), the LULC features were defined with their corresponding landuse capability classes, as listed in Table 4. For instance, the areas comprised of irrigated croplands are designated as having lower sensitivity to degradation. Similarly, natural land cover (i.e., forests, plantations) is generally categorized as having low sensitivity, except in areas with shrub cover, which are considered to be at a higher risk.

Table 4.

Land use capability classes as per the NBSS & LUP (Singh et al., 2016)⁸³.

Land capability class	Description
II	Good-quality cultivable lands have only minor constraints associated with soil texture, drainage, and susceptibility to erosion.
III	Lands deemed moderately good for cultivation exhibit moderate problems related to erosion, slope, drainage, and soil characteristics.
IV	Fairly good cultivable lands experience severe limitations in erosion, drainage, climate, and soil properties.
V	Soils with minimal or no erosion hazards still possess other inherent limitations that are difficult to correct, restricting their use largely to pasture, rangeland, forestland, or wildlife food and habitat.
VII	Soils characterized by severe limitations are generally unsuitable for cultivation and are typically employed for grazing, forestry, or wildlife management.
VIII	Lands suitable only for wildlife, recreation, and quarrying.
Miscellaneous	Water bodies, Settlements/built-up area

The MQI assesses the management quality of LULC features by considering inherent soil characteristics, climatic conditions, and environmental factors that affect sustainable land productivity. Figure 9(a) illustrates various LULC features, including significant areas of agricultural land, barren land, fallow land, land with shrub cover, upland forests, foothill plantations, rocky outcrops, mines and quarries, stony wastelands, water bodies (such as rivers, channels, and lakes), as well as urban and rural settlements. The LULC features in alluvial plains are noted for poor land management practices, placing these areas at an increased risk of erosion and salinity. Well-managed upland forest areas, foothill plantations, and riverbed croplands demonstrate reduced sensitivity to land degradation, particularly when climatic conditions are favorable. Irrigated croplands are similarly classified as low-risk due to their adherence to proper management practices. Figure 9(b) illustrates the land use capacity linked to various LULC features, reflecting their production potential under stress. The map categorizes the land into six classes. Classes II, III, and VIII show higher production capacity, mainly in forests, plantations, and irrigated croplands. Moderate production capacity is observed in Classes III and IV, commonly in foothills, alluvial plains, and areas with shrub cover. Classes V and VII, comprising fallows, unirrigated croplands, barren lands, and eroded areas, exhibit lower production potential and are highly vulnerable to degradation. Settlements (class I) and water bodies (class VI) are excluded from the analysis, as they do not influence land capability.

Fig. 9.

Landuse Management Quality Index (MQI) parameters: (a) Landuse/land cover (LULC), (b) Land use capability classes. The maps were generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Figure 10 illustrates the spatial distribution of the MQI based on their LULC classes, indicating the spatial peculiarities of land productivity in the district. The MQI emphasizes the land potential of areas influenced by both natural variables and human activity (Poornazari et al., 2021). Within the district, land use and land cover (LULC) research indicates that a substantial area of 1,583.78 km², accounting for 55.20%, is characterized by intermediate land use management quality. Wherein, the MQI value ranges from 1.25 to 1.50. This region encompasses a diverse range of land types, including alluvial plains, irrigated croplands, barren lands, fallows, shrublands, and unirrigated drylands, classified within land capability classes II and III, which signifies a higher risk of deterioration. Moreover, 29.73% of the district displays a better rate of land management quality, with an MQI of 1.00 to 1.25. This area is primarily composed of upland forests, plantations, and irrigated croplands, which are capable of delivering good productivity throughout the year and are less prone to land and soil degradation (Table 5). Conversely, 11.13% of the district is characterized as having lower land management quality, with an MQI greater than 1.50. This comprises foothill slopes, barren plains, fallows, shrublands, degraded valley lands, and salt-affected places, all of which are very sensitive to degradation. Nearly half of the overall district area, mainly cultivable fields, falls under the intermediate land use management quality category, raising concerns about their vulnerability to degradation, likely due to insufficient land use management approaches.

Fig. 10.

Land Use Management Quality Index (MQI) map of the study area. The map was generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Table 5.

Quantitative summary of quality indices.

Parameter	Class	Index	Area (sq. km)	Area (%)
SQI	High	< 1.13	39.90	1.39
	Moderate	1.13–1.45	1883.27	65.64
	Low	> 1.45	833.01	29.03
CQI	High	1–1.25	950.70	33.14
	Moderate	1.25–1.50	731.15	25.48
	Low	> 1.50	1074.33	37.45
VQI	High	< 1.13	862.89	30.08
	Moderate	1.13–1.38	836.11	29.14
	Low	> 1.38	1057.19	36.85
MQI	High	1–1.25	853.04	29.73
	Moderate	1.25–1.50	1583.78	55.20
	Low	> 1.50	319.36	11.13

Land degradation vulnerability index (LDVI) assessment and mapping

The Land Degradation Vulnerability Index (LDVI) evaluates the status of four key indices at each site: the Soil Quality Index (SQI), Climate Quality Index (CQI), Vegetation Quality Index (VQI), and Management Quality Index (MQI). Figure 11 presents the LDVI map, classifying areas according to their vulnerability to land degradation. The classification includes: (i) non-affected zones (N) with LDVI values below 1.17; (ii) Potential vegetative cover (P) for values between 1.17 and 1.22; (iii) Fragile zones, subdivided into F1, F2, and F3, with values ranging from 1.23 to 1.37; and (iv) Critical zones, subdivided into C1, C2, and C3, for values above 1.38. Among these, the C3 category represents the most critical areas, exhibiting LDVI values greater than 1.53, which indicates a severe risk of land degradation. Table 6 summarizes the LDVI classes and their spatial distributions, providing a comprehensive overview of areas in the Theni district that are sensitive to land degradation. The statistical measures of the LDVI classes and their areal extent are displayed in Fig. 12. Approximately 25.38% of the land in the district is critically susceptible to degradation, with severe degradation primarily occurring in the northern tip, specifically in the Periakulam region and around the Vaigai reservoir, classified under the C1, C2, and C3 classes. Very critical regions (C3) are predominantly located in the alluvial plains of the central and northeastern parts of Theni district. Natural factors, including steep slopes and inadequate land management practices, have exacerbated land degradation and soil loss in these areas. The spatial distribution of highly degraded land is quantified as 291.06 km² (10.15%) for category C1, 331.42 km² (11.55%) for subclass C2, and 105.53 km² (3.68%) for subclass C3. Land use and land cover (LULC) in these critical zones are highly vulnerable to degradation, primarily due to factors such as soil erosion, salinity intrusion, nutrient depletion, and insufficient management interventions.

Fig. 11.

The Land Degradation Vulnerability Index (LDVI) map illustrates the spatial distribution of land degradation zones. The map was generated in ArcMap v10.6 software (https://www.esri.com/en-us/arcgis/products/arcgis-desktop/overview).

Table 6.

LDVI classes and their area extend in the study area.

Class	Subclass	LDVI value	Area (km2)	Area (%)
Non affected	N	< 1.17	381.81	13.31
Potential	P	1.17–1.22	372.77	12.99
Fragile	F1	1.23–1.26	172.22	6
	F2	1.27–132	480.96	16.76
	F3	1.33–1.37	620.4	21.62
Critical	C1	1.38–1.41	291.06	10.15
	C2	1.42–1.53	331.42	11.55
	C3	> 1.53	105.53	3.68

Fig. 12.

Statistical measurement of the LDVI classes with the areal extent.

Fragile zones across the district demonstrate considerable vulnerability to degradation and are strongly linked to adjacent critical zones. In these fragile areas, LULC features, including croplands, fallows, barren lands, and eroded regions, undergo periodic changes driven by rainfall patterns, land surface temperature (LST), soil moisture levels, climate variability, and human-induced stresses. LULC features are heavily influenced by land and soil degradation processes within the critical zones, leading to substantial declines in agricultural productivity. Numerous sections of land, including croplands, fallows, barrens, and degraded lands, undergo considerable degradation, resulting in widespread desertification. Approximately 44.38% of the land is classed as fragile to deterioration, with 21.62% labelled as severely fragile (F3). Cultivable fields in regions like Bodinayakanur, Veerapandi, and the western parts of Rajagopalanpatti are particularly vulnerable. Intensive land use and poor methods that disregard the land’s carrying capacity have resulted in serious land degradation, affecting both these places and their surroundings. An area of around 480.90 km² (16.76%) is classified as having F2 fragility to deterioration, while an additional 172.22 km² (6%) comes under F1 fragility. These zones encompass land use along the foothills and croplands adjacent to metropolitan areas. The significant areas affected by these classifications are the Cumbum and Uthamapalayam blocks, which are recognized for their substantial agricultural operations. The LULC features inside fragile zones have negatively impacted agricultural productivity, as indicated by various sections of croplands, fallows, barrens, and degraded lands. Consequently, the LULC in these places is subject to degradation due to soil-water stress and numerous human stresses.

Notably, an area of 372.77 km² (12.99%) is defined as potentially vegetative cover (P), whereas around 381.81 km² (13.31%) is designated as a non-affected zone. This latter category comprises upland forest cover, plantations, and irrigated croplands. The LULC features within these zones demonstrate enhanced biomass productivity, supported by adequate rainfall, sufficient soil moisture, and favourable soil fertility. These areas also exhibit higher resilience to land and soil degradation throughout all seasons. In contrast, irrigated and unirrigated croplands, fallows, and other regions subjected to poor land management practices exhibit heightened sensitivity to degradation, often transitioning into eroded or saline-affected soils. The primary factors contributing to land degradation include spatio-temporal variability in rainfall, fluctuations in soil moisture, saline intrusion, and anthropogenic pressures, all of which drive severe soil and land deterioration across different parts of the district.

Spatial correlation of LDVI and quality indices

The spatial correlation between the LDVI and Quality Indicators (QIs) was assessed using a spatial statistics tool in GIS software. Using the attribute data, Pearson’s correlation coefficient was calculated to estimate the degree of correlation among the Quality Indicators (QIs). Table 7 presents a summary of the Pearson correlation coefficients between LDVI and Quality Indicators (QI), indicating that these QI variables significantly influence LDVI at a site-specific scale and exhibit spatial relationships (i.e., positive or negative correlations) among themselves. Figure 13 illustrates the spatial correlations between LDVI and the Quality Indicator using scatter plots. The results demonstrate the significance of the quality indicators (SQI, MQI, VQI, and CQI) in relation to the LDVI at the site-specific scale, reflecting their association with geo-environmental variables that induce land degradation across areas experiencing changing LULC and climate conditions.

Table 7.

Pearson correlation coefficients of land degradation vulnerability index (LDVI) and quality indicators.

Indicators	LDVI	VQI	CQI	MQI	SQI
LDVI	1	0.84	− 0.75	0.83	0.99
VQI	0.84	1	− 0.65	0.67	0.82
CQI	− 0.75	− 0.65	1	− 0.58	− 0.74
MQI	0.83	0.67	− 0.58	1	0.83
SQI	0.99	0.82	− 0.74	0.83	1

Significant values are in bold.

Fig. 13.

Scatterplots showing the spatial correlations of LDVI with Quality Indicator: (a) LDVI vs. SQI, (b) LDVI vs. VQI, (c) LDVI vs. CQI, and (d) LDVI vs. MQI.

The findings showed that SQI has the highest positive correlation with the LDVI, with a correlation coefficient (r) value of 0.99, emphasizing the importance of soil properties and health in mitigating land degradation. Moreover, it showed a direct relationship between SOI and LDVI; areas with good soil quality and proper implementation practices have retained soil health and productivity, and these areas are characterized by a lower LDVI rate, especially in riverbank agricultural lands and irrigated croplands. Secondly, the VQI and MQI showed a high positive correlation with LDVI, with correlation coefficient values of 0.84 and 0.83, respectively. These Quality Indicators (SQI, VQI, and MQI) significantly influenced the LDVI at various points, underscoring their importance in mitigating land degradation risks through sustainable management practices. The areas with a higher quality index (QI) of these indicators experienced good health of vegetation and crop productivity. At the same time, those with lower LDVI values are less sensitive to erosion, runoff, soil infertility, and salinity ingress. Overall, the findings revealed a positive correlation between SQI and MQI, with a coefficient (r) of 0.83, demonstrating their interdependence in minimizing land degradation risk across the study area.

Conversely, the CQI displays a negative correlation with the LDVI, with a coefficient (r = -0.74), indicating an inverse relation with climate and land degradation risk, where areas are affected by drastic land degradation vulnerability, while facing extremely high or low rainfall, temperature, evapotranspiration, and altering physico-chemical and morphological properties of soil-land-water, causing severe impacts on vegetation health. Areas with lower CQI, characterized by excessive aridity and sparse rainfall, experience a significant increase in vulnerability to land degradation, primarily in unirrigated croplands and fallow lands. This frequently results in climate constraints that accelerate soil erosion and degrade cropland and vegetative cover. In contrast, CQI and VQI are considered more sensitive indicators, exhibiting a negative correlation with LDVI due to their inverse relationship with other spatial parameters. Wherein the CQI has a negative correlation with LDVI (-0.75), VQI (-0.65), MQI (-0.58), and SQI (-0.74), indicating changes (increase or decrease) in CQI variables (i.e., rainfall, aridity, land surface temperature and evapo-transpiration), inducing land degradation risks, and leading to adverse impacts on land-water-soil and vegetation properties. Therefore, the variables of CQI and VQI are required for monitoring through proper management approaches to reduce their impact on degradation risk.

Validation of LDVI output using ROC-AUC method

The Area Under Curve (ROC-AUC) is employed to evaluate the relative accuracy of the LDVI output in comparison with NDVI (i.e., pixel-based assessment of vegetation health). The AUC value reflects the model’s performance, ranging from 0.5 to 1; therefore, a higher AUC (> 0.8) indicates strong predictive ability, whereas values (< 0.5) suggest random performance. The ArcSDM tool was used to determine the AUC value by calculating the spatial correlation coefficient between the LDVI and NDVI, which assesses the model’s capacity to distinguish between positive and negative cases within the dataset. Figure 14 illustrates the AUC-ROC curve of LDVI and NDVI, derived from point-cloud data associated with respective raster layers. The graph displays the True Positive Rate (sensitivity) versus the False Positive Rate, with sensitivity, i.e., true positive rate, shown on the y-axis and the false positive rate on the x-axis. The findings demonstrate a significant spatial correlation between LDVI and NDVI, where elevated LDVI values correlate with diminished NDVI values, indicating specific sites (pixel size of 30 × 30 m) with limited vegetation that align with crucial degradation zones. The predicted AUC value is 0.83, indicating a dependable performance (83%) of the LDVI output compared with site-specific NDVI values. Wherein the NDVI reflects the occurrence of plants with greenness and health over time, enabling the identification of areas under land degradation, based on vegetation-moisture stress, removal of vegetation, and extend of non-vegetation areas.

Fig. 14.

AUC curve-based cross-validation of the LDVI map.

The LDVI-NDVI comparative studies revealed that the areas with an NDVI value (> 0.51) imply healthy vegetation zones (i.e., forest cover, plantations, and irrigated croplands), indicating the lower LDVI rate (< 1.22), which comes under non-affected (N) zones, and potentially vegetated zone (P). Whereas the LDVI value (> 1.38) directly matched with the areas having lower NDVI (0.1–0.2), which reflects the occurrence of moisture-stressed plants, fallows, barren land, shrublands, and other degraded landforms. Notably, the NDVI value (0.27–0.5) relies with the LDVI (1.33–1.37), indicating fragile zones, with stressed vegetation, particularly in the middle and northeastern alluvial plains. Importantly, the NDVI values (< 0.1) indicate non-vegetative features, such as built-up regions and water bodies, which are masked in the LDVI map. The site-specific relationship between LDVI and NDVI demonstrates satisfactory accuracy; however, its reliability needs to be improved by incorporating field sampling data on soil nutrients and physico-chemical properties, which will enhance the on-site reliability of the LDVI output.

Conclusion

The study’s outcome supports the United Nations Sustainable Development Goals (UN-SDGs 13 – Climate Action, and SDG 15 – Life on Land) by addressing the risk of land degradation and the impacts of climate change on soil-water-land properties in a vulnerable agro-climatic zone. The LDVI map illustrates the spatial characteristics of land degradation risk at a specific scale within the western agro-climatic zone of the Theni District, located in the southern Indian state of Tamil Nadu. The generated LDVI map indicates that 25.38% of the area is in a critical zone for land degradation. The analysis of the Quality Indices indicates that soil quality (SQI) and management quality (MQI) exert the most significant influence on land degradation in the Theni District, as reflected by their strong positive correlations with LDVI. Vegetation quality (VQI) also exhibits a notable association (r = 0.84), underscoring the importance of maintaining vegetation cover in mitigating erosion and water stress. The LULC features, i.e., croplands, fallows, barren fields, and eroded areas, are commonly found in the central alluvial plains, western foothill ranges, and northeastern plains. These areas are particularly vulnerable to land degradation due to factors such as erosion, nutrient depletion, and poor land management practices. Conversely, only 13.31% of the area is classified as unaffected by degradation, while 12.99% is marked by potential vegetative cover, demonstrating greater resilience to land degradation throughout all seasons. The findings suggest that the risks associated with land degradation in vulnerable LULC areas can be mitigated through effective management practices for land, soil, and water resources. Overall, this study provides a geo-database on land, soil, and water, which is valuable for the sustainable management of land resources in the long term. The current LDVI assessment, although effective in identifying sensitive zones of degradation, has certain limitations. The accuracy of the results is influenced by the spatial resolution and temporal coverage of the input datasets, particularly the climatic and soil parameters, which were derived from secondary sources. The lack of field-based validation data, such as soil moisture, organic matter, and erosion rates, restricts the ability to calibrate the model outputs. Future studies incorporating in-situ observations, such as soil chemical properties, soil salinity, groundwater quality, and groundwater recharge, as well as high-resolution time-series imagery and multi-temporal monitoring, could enhance precision and temporal responsiveness to support informed decision-making in scientific contexts.

Author contributions

S. Kaliraj: research framework, original draft review and editing, visualization, supervision; R. J. Jerin Joe : Writing review and editing of the original draft; V Stephen Pitchaimani and S. Richard Abishek : contributed to the review and editing of the manuscript. Shankar Karuppannan : Formal analysis, Writing- Reviewing and Editing. All authors have read and agreed to the published version of the manuscript.

Funding

The authors received no funding for this work.

Data availability

The data that support the findings of this study are available on request from the corresponding author.

Declarations

Competing interests

The authors declare no competing interests.

Human face consent to publish a declaration

We have carefully reviewed all the images in our manuscript and confirm that no human faces are present.