Deepti Soni*
Samina Yasmin
Aman Chandrakar
Anju Jangade
Understanding the spatial and temporal distribution of soil moisture is crucial for sustainable agricultural practices, hydrological modelling, and climate-resilient land management. With increasing pressure on water resources and changing land use dynamics, especially in agrarian districts like Dhamtari, there is a growing need for reliable, spatially continuous methods to monitor soil moisture variability. Traditional approaches often fall short in scale and frequency, prompting the adoption of remote sensing-based methods that leverage satellite data to estimate key environmental indicators. Among these indicators, the Soil Moisture Index (SMI) and Land Surface Temperature (LST), derived from remote sensing, offer significant insights into the dynamic relationship between soil moisture, land surface characteristics, and temperature patterns. The integration of these two indices enables comprehensive assessment of soil moisture variations across time and space. In recent years, the Temperature-Vegetation Dryness Index (TVDI) has also emerged as a powerful tool for evaluating the vegetation condition and its relationship with moisture availability. TVDI combines both temperature and vegetation data, making it highly effective in understanding vegetation stress and moisture depletion across different landscapes. By examining the TVDI, SMI, and LST together, we can obtain a multifaceted perspective of land surface dynamics, providing more accurate assessments of drought, soil health, and vegetation conditions. This report aims to analyze the temporal and seasonal variations of SMI, LST, and TVDI across Dhamtari district from 2021 to 2024, focusing on how these indices interact and influence land surface conditions. By integrating these indices, this study will explore key environmental patterns, correlations, and trends to enhance our understanding of soil moisture distribution, vegetation stress, and temperature variations, all of which play a critical role in shaping the region’s agricultural and environmental strategies.
2. Software and Tools Used
-
- QGIS (v3.28): Used for NDVI, LST, and SMI computations, and map rendering
- Google Earth Engine: Time-series visualization and data pre-processing
- USGS Earth Explorer: Satellite imagery download
- Microsoft Excel: Data integration and tabulation
3. METHODOLOGY
In order to meet the objectives, the following research methodology was adopted:
3.1 Study Design & Planning
Data Collection
- Field surveys for spot soil moisture (2025)
- Landsat-8 OLI/TIRS satellite imagery (2021–2024) for January, May, and October
- Administrative boundary data (Bhuvan)
- Land Use/Land Cover data (NRSC)
- Population and census datasets
- Software and tools used
3.2.1 QGIS (v3.28): Used for NDVI, LST, and SMI computations, and map rendering
3.2.2 Google Earth Engine: Time-series visualization and data pre-processing
3.2.3 USGS Earth Explorer: Satellite imagery download
3.2.4 Microsoft Excel: Data integration and tabulation
3.3 Soil Moisture Index (SMI) Estimation
3.4 Spatio-temporal Analysis
- Seasonal comparisons: January, May, and October
- Year-wise analysis: 2021–2024
- Moisture classification into five distinct categories
3.5 Correlation & Statistical Analysis
- SMI vs. LST
- SMI vs. TVDI
- SMI vs. NDVI
- Pearson correlation coefficient analysis
3.6 Interpretation & Visualization
- Map generation for LST, SMI, TVDI and NDVI
- Identification of drought-prone and moisture-deficient regions
3.7 Findings and Discussion
Results and Discussion
This section presents an integrated interpretation of the observed patterns in Land Surface Temperature (LST), Soil Moisture Index (SMI), Normalized Vegetation Index (NDVI) and Temperature Vegetation Dryness Index (TVDI) across spatial and temporal dimensions in Dhamtari district from 2021 to 2024. By analyzing these indices seasonally (January, May, and October), temporally (year-on-year), and through correlation, a comprehensive understanding of land surface dynamics, vegetation stress, and soil moisture fluctuations is derived. Findings are structured to reveal critical insights into climate stress responses, drought This chapter consolidates and interprets the analytical outcomes derived from the temporal, seasonal, and correlation analyses of Land Surface Temperature (LST), Soil Moisture Index (SMI), Temperature Vegetation Dryness Index (TVDI), and Normalized Difference Vegetation Index (NDVI) across Dhamtari district from 2021 to 2024. Drawing from spatial datasets and remote sensing indices analyzed for January, May, and October of each year, this section elaborates on the inter-seasonal variability, inter-annual fluctuations, and inter-variable relationships, offering insight into patterns of thermal stress, moisture availability, vegetation condition, and drought susceptibility. Tables and maps, referenced throughout, provide a visual and quantitative basis for understanding the spatio-temporal dynamics and how vegetation health (as indicated by NDVI) responds to changes in moisture and temperature regimes. Behavior, and their implications for sustainable land and water management.
5. Year-wise Correlation Results -
Table 5. 1 Pearson Correlation Coefficient between Variables
|
Year |
Month |
SMI–LST (r) |
SMI–TVDI (r) |
SMI-NDVI |
|
2021 |
Jan |
–0.62 |
–0.85 |
0.88 |
|
May |
–0.72 |
–0.89 |
0.79 |
|
|
Oct |
–0.60 |
–0.78 |
0.94 |
|
|
2022 |
Jan |
–0.59 |
–0.81 |
0.86 |
|
May |
–0.77 |
–0.91 |
0.76 |
|
|
Oct |
–0.51 |
–0.76 |
0.95 |
|
|
2023 |
Jan |
–0.64 |
–0.88 |
0.87 |
|
May |
–0.80 |
–0.92 |
0.78 |
|
|
Oct |
–0.53 |
–0.73 |
0.96 |
|
|
2024 |
Jan |
–0.55 |
–0.83 |
0.89 |
|
May |
–0.74 |
–0.87 |
0.81 |
|
|
Oct |
–0.49 |
–0.69 |
0.97 |
CONCLUSION
- mixed results, varying by year depending on monsoon performance. The comprehensive spatial-temporal assessment of Soil Moisture Index (SMI), Land Surface Temperature (LST), Temperature Vegetation Dryness Index (TVDI), and Normalized Difference Vegetation Index (NDVI) across Dhamtari district from 2021 to 2024 provides critical insights into the hydro-climatic and ecological dynamics of the region. The analysis, conducted across three key months (January, May, and October), reveals a recurring pattern of seasonal stress and partial recovery, with May emerging as the most hydro-thermally stressed period in all years.
- LST reached its maximum values in May, with temperatures exceeding 42°C in 2023 and 2024, especially in the central-western zone (Map 1.6, Map 1.9). This thermal intensification was directly linked to the steep decline in SMI, particularly in May 2022 and 2023 when over 60% of the district fell under Class 1 and 2 (SMI < 0.4) indicating acute surface dryness (Table 2.5). January consistently exhibited higher SMI, reflecting cooler conditions and greater soil moisture retention, while October showed
- TVDI values mirrored the trends in LST and SMI, with May showing peak dryness, where more than 2,000 sq.km of land in 2022 and 2023 registered TVDI > 0.8 (Map 2.12, Map 2.15). January values remained mostly below 0.4, indicating minimal drought stress during winter.
NDVI trends added a crucial ecological layer to the analysis. The lowest NDVI values (<0.2) were observed during May in all years, reflecting poor vegetation cover and high stress on crop and forest systems (Map 5.38, Map 5.41). January and October, by contrast, showed higher NDVI values in Class 3 and 4 (0.2–0.6), signifying partial to moderate greenness recovery (Table 5.10). NDVI not only helped spatially map vegetation condition but also responded effectively to changes in both LST and soil moisture availability.
REFERENCE
- Ahmad,?I.,?et?al. (2021). Utilizing TVDI and NDWI to Classify Severity of Agricultural Drought in Malaysia. Agronomy, 11(6):1243. Provides methodology on combining NDVI and LST within TVDI framework. [1]
- Barrett,?A.B.,?Duivenvoorden,?S., et?al. (2019). Forecasting vegetation condition for drought early warning systems in Kenya. [2]
- Chen, Z.,?Chen,?X.,?et?al. (2023). Temperature Vegetation Drought Index for Drought Monitoring in Guangdong, China. Remote Sensing, 15(9). Shows TVDINDVI’s stronger drought response than TVDIEVI?pubmed.ncbi.nlm.nih.gov
- Dennison,?P.E.,?Roberts,?D.A.,?et?al. (2005). NDWI for monitoring vegetation moisture content. Int J Remote Sens, 26(5):1035–1042?en.wikipedia.org [4]
- Dlamini,?Z.,?et?al. (2022). Monitoring of Vegetation Drought Index based on Landsat in China. Applied Sciences, 14(19):8904. Utilizes TVDI with NDVI–LST space for regional drought assessment.
- Du,?J.,?et?al. (2012). Improved TVDI (TVDIm) for drought monitoring in semi-arid China. MDPI, validated with SPEI and precipitation correlation. [6]
- Elhaddad,?M.,?Elhag,?M.,?Abdelrahman,?M. (2022). Soil moisture estimation using LST in Sudan savanna region. IRJET, 9(12):1566–1570.?
- Farg,?E.,?et?al. (2005). Use of NDWI for monitoring live fuel moisture. Int J Remote Sensing, 26(5):1035–1042?en.wikipedia.org
- Gao,?B.-C. (1996). NDWI – Normalized Difference Water Index for vegetation water content. Remote Sensing of Environment, 74(3):570–581?en.wikipedia.org+1mdpi.com+1 [9]
- Holben,?B.N. (1986). Maximum-value composite images from AVHRR data. Int J Remote Sensing, 7(11):1417–1434?en.wikipedia.org [10]
- Jiang, L., Wang,?P., Jiang,?L., Gong,?L., Li,?X., Zhang,?H., Wang,?X. (2021). Temperature Vegetation Dryness Index and Its Application in Agricultural Drought Monitoring. Chinese Agricultural Science Bulletin, 37(29):132–139?[11]
- Kshetri,?R.,?Dahal,?D. (2024). Estimation of LST & SMI using satellite imagery – Nepal case study. Global Journal of Human-Social Science Research, 24(2). [12]
- Mortier,?S.,?Hamedpour,?A.,?et?al. (2023). Inferring soil temperature–NDVI relationship via machine learning. ?arxiv.org
- Nay,?J.J.,?Burchfield,?E.,?Gilligan,?J. (2016). Machine learning for forecasting remotely sensed vegetation health. ?arxiv.org [14]
- Patel,?K.,?Dadhania,?D. (2021). Mapping of Soil Moisture Index using Remote Sensing & GIS – Saurashtra, India. Journal of Remote Sensing Applications, 7(1):22.
- Qihao Weng (2004). Estimation of land surface temperature–vegetation relationship for urban heat island studies. Remote Sensing of Environment, 89(4):467–483?[16]
- Raj,?R.,?Deshmukh,?P.,?Rane,?A. (2020). Analysis of Soil Moisture using LST in Semi-Arid India. Indian Journal of Geo-Marine Sciences, 49(11):1814–1821.?
- Sahana,?M.,?Ghosh,?S.,?Panda,?R.K. (2012). Drought assessment using Remote Sensing & GIS – West Bengal, India. Int J Remote Sensing, 33(8):2463–2483.
- Sánchez,?N.,?González-Zamora,?A.,?et?al. (2020). Integrated SMOS & MODIS drought index (SMADI) global agricultural monitoring. ?arxiv.org [19].
10.5281/zenodo.15735323