Seasonal drought assessment using the Normalized Difference Drought Index (NDDI) in Satara District of Western Maharashtra (2014-2023)

Introduction

Drought is a significant detrimental climatic phenomenon for ecosystems and human society. Recent years have seen a sharp rise in worldwide drought episodes as a result of accelerated climate change. For the purposes of water management and drought mitigation, precise and reliable worldwide mapping and monitoring of drought severity is required. Regional drought assessment and mitigation initiatives should benefit from global products obtained from operational satellite remote sensing, particularly in regions of the world where alternative information sources are limited by inadequate infrastructure development and sparse measurement networks (Mu et al., 2013) [8] . Agriculture forms the backbone of the Indian economy and is often the first sector to feel the effects of drought, owing to its heavy dependence on soil moisture reserves and water availability during various stages of crop growth. Planning to lessen the effects of drought would be made easier with an understanding and development of technologies for drought monitoring and prediction. There are several methods for describing drought, however the use of drought indices is common (Tsakiris et al., 2007) [10] . When comparing the different indices, Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI) and Normalized Difference Drought Index (NDDI) are better suited for tracking the characteristics of agricultural drought (Stancalie and Serban, 2014) [9] . The NDDI is thought to be more accurate than the NDVI or NDWI algorithms alone for identifying drought on agricultural land. It has a higher sensitivity level to drought due to computations from the NDVI and NDWI modifications (Dobri et al., 2021) [5] . The main goal of the NDDI is to record the spatio-temporal pattern and severity of the drought by integrating multiple indicators to investigate the phenomenon. As an indicator of drought conditions, the NDDI value ranges from -1 to 1, with higher values denoting a more severe drought. The NDVI is sensitive to vegetation, while the NDWI is sensitive to water resources and water content in leaves. The NDDI uses these two facts in their drought prediction. (Khampera et al., 2018). Satara district in western Maharashtra presents distinct agroclimatic contrasts, with high-rainfall areas in the west and drought-prone conditions toward the east. This spatial variability, together with seasonal cropping practices, makes it a suitable region for monitoring drought dynamics over time. This study aims to conduct a spatio-temporal analysis of NDDI from 2014 to 2023 for Satara district, assessing seasonal variations (rabi, summer and kharif) and their relationship with rainfall. The results will highlight regions vulnerable to drought and provide useful information for sustainable agricultural planning and resource management in the district

Materials and Methods

Study area

Satara district lies in the western region of Maharashtra, extending between 17°05′ to 18°11′ N latitude and 73°33′ to 74°54′ E longitude as shown in the Fig1. Rainfall distribution in the district shows a strong west-east gradient, with Mahabaleshwar in the western side receiving more than 6000 mm of annual rainfall, while eastern tahsils such as Man, Phaltan, Khandala, and Khatav receive only about 500 mm or even less (Barakade, 2012) [2] . The eastern region is therefore highly susceptible to drought conditions. As noted by Pravin and Hemalata (2020), among the 11 tahsils, Khandala, Phaltan, Man and Khatav experience severe drought and Koregaon shows partial drought influence.

Normalized Difference Drought Index

The Normalized Difference Drought Index algorithm is considered as more accurate for detecting drought on agricultural land compared to using the NDVI or NDWI algorithms individually. This is because it combines the results of the NDVI and the NDWI, which enhances its sensitivity to drought conditions (Dobri et al., 2021) [5] . NDDI is the ratio of difference between NDVI and NDWI to the sum of NDVI and NDWI bands, it expressed as: NDDI = (NDVI -NDWI) / (NDVI +NDWI) Where, NDVI is calculated as (NIR -RED) / (NIR + RED) with NIR and RED indicating near-infrared and red reflectance, respectively (Deering, 1978) [4] . The NDWI is derived as (NIR -SWIR) / (NIR + SWIR), where NIR and SWIR are the reflectance in the near infrared and short wave infrared bands (Gao, 1996) [6] . NDDI is considered more sensitive to drought conditions than the NDVI and NDWI. A high NDDI value indicates drought, typically corresponding with low NDVI and NDWI values, reflecting reduced vegetation and moisture. Conversely, a low NDDI value suggests no drought or water scarcity, with NDVI and NDWI values being relatively high, indicating healthy vegetation and adequate water availability. The NDDI mainly divided into five classes (Artikanur and Setiawan (2022) [1] , Herawati et al. (2024)) [7] . These classes are given in the Table 1.

Results

Season-wise Temporal Analysis of Mean NDDI (2014-2023)

Seasonal analysis of mean NDDI from 2014 to 2023 was carried out to assess drought variability across rabi, summer and kharif seasons in Satara district which is presented in Table 2 and depicted in the Fig 2 . NDDI reflects both vegetation health and moisture stress, and its seasonal trends help in identifying drought-prone 3 and illustrated in Figure 6. From the Table 3 and Figure 6, during the rabi season in Satara district, the NDDI reflected a consistent dominance of low drought severity areas over the ten-year period.

Summer Season

The area-wise drought severity classification using NDDI during the summer season (2014-2023) is presented in Table 4 and illustrated in Figure 7.

Kharif Season

The area-wise drought severity classification using NDVI during the kharif season (2014-2023) is presented in Table 5 and illustrated in Figure 8. This reflects favourable rainfall distribution during the monsoon, supporting healthy vegetation growth. In contrast, both rabi and summer seasons exhibited relatively higher drought severity, with notable shares of area under moderate to high severity categories. Summer season showed the greatest drought intensity, with a larger area affected by high and very high severity, especially in 2016 and 2019, attributed to higher evapotranspiration demand and minimal rainfall. These results align with the findings of Davaagerel et al (2021) [3] who reported highest drought severity in summer season. The overall trend emphasizes the dependence of vegetation on monsoonal rainfall and the necessity of efficient water management during the drier seasons. Rainfall and NDDI exhibited a negative association across all three seasons (Table 7), indicating that drought severity (higher NDDI) increases as rainfall decreases. The negative correlation between NDDI and rainfall across seasons concurs with Dobri et al. (2021) [5] , reinforcing the index's utility in drought monitoring. The strongest and statistically significant correlation was observed during kharif (r = -0.70, p < 0.05), highlighting the sensitivity of vegetation moisture stress to rainfall variability in the monsoon driven season. A moderate negative correlation during summer (r = -0.56, p < 0.10) indicates partial rainfall control on vegetation dryness when irrigation or residual soil moisture may buffer drought effects. Conversely, rabi showed a very weak and insignificant relationship (r = -0.22, p > 0.10), likely due to limited rainfall and agricultural dependence on irrigation during this season.

Summary

The study investigated the spatio-temporal drought patterns in Satara district of Western Maharashtra using the Normalized Difference Drought Index (NDDI) derived from Landsat-8 Collection 2 Level-2 data for a ten-year period (2014-2023). NDVI and NDWI were computed for each season (rabi, summer and kharif), and integrated to generate NDDI maps and seasonal drought severity classifications. The study also examined the relationship between NDDI and seasonal rainfall. Season-wise temporal analysis revealed clear differences in drought intensity. Summer consistently exhibited the highest NDDI values, reflecting severe moisture stress due to high temperatures and minimal rainfall. Rabi season showed moderate drought intensity with considerable year-to-year variation, while kharif season exhibited the lowest NDDI values owing to monsoonal rainfall and favourable vegetation conditions. Spatial analysis of NDDI maps indicated that eastern parts of Satara including Man, Khatav, Phaltan, and Khandala experienced persistent drought across years and seasons. In contrast, western regions near Mahabaleshwar consistently showed lower drought severity because of higher rainfall and denser vegetation cover. In rabi, 2014 recorded high drought intensity, whereas 2017 reflected better conditions. During summer, 2014 showed the most severe drought conditions, while 2023 demonstrated notable improvement. kharif seasons generally showed lower drought severity, with 2015 as a drier year and 2019 reflecting healthier vegetation cover. Correlation analysis between rainfall and NDDI revealed a significant negative relationship during kharif (r = -0.70), indicating that rainfall strongly controls vegetation moisture conditions in this season. Summer showed moderate negative correlation (r = -0.56), while rabi exhibited weak association because most crops rely on irrigation rather than rainfall. Overall, the NDDI proved highly effective in identifying spatial and seasonal drought variability and highlighting drought-prone regions in Satara district.

Conclusion

The spatio-temporal evaluation of NDDI from 2014-2023 demonstrates clear seasonal drought dynamics in Satara district. Summer emerged as the most drought-prone period due to high temperatures and minimal rainfall, whereas kharif showed minimal drought stress due to reliable monsoon rainfall. Rabi season exhibited moderate drought severity with stable vegetation conditions supported by irrigation practices. Persistent drought hotspots in eastern tahsils highlight their heightened vulnerability and demand priority attention in future water resource planning. The significant negative relationship between kharif rainfall and NDDI highlights rainfall dependency of the region's agriculture. The analysis further reveals that NDDI is an effective indicator for assessing drought, as reflected by its strong negative correlation with rainfall, particularly during kharif. Overall, the findings emphasize the importance of continuous drought monitoring, strategic water management and targeted mitigation efforts to support sustainable agriculture in the drought sensitive regions of Satara district.

References

1. Artikanur SD, Setiawan Y. Normalized difference drought index (NDDI) computation for mapping drought severity in Bojonegoro Regency, East Java, Indonesia. IOP Conf Ser Earth Environ Sci. 2022;1109(1):012027.
2. Barakade AJ. Growth of population in Satara district of Maharashtra. World Res J Geol. 2012;1(2):17-20.
3. Davaagerel M, Munkhchuluun B, Bayarsaikhan U. Assessment of drought severity in Khar Yamaat Nature Reserve Area using satellite derived drought indexes. Mong J Biol Sci. 2021;19(2):61-70.
4. Deering DW. Rangeland Reflectance Characteristics Measured by Aircraft and Spacecraft Sensors. College Station (TX): Texas A&M University; 1978.
5. Dobri RV, Sfîcă L, Amihăesei VA, Apostol L, Țîmpu S. Drought extent and severity on arable lands in Romania derived from normalized difference drought index (20012020). Remote Sens. 2021;13(8):1478.
6. Gao BC. NDWI-A normalized difference water index for remote sensing of vegetation liquid water from space. Remote Sens Environ. 1996;58(3):257-266.
7. Herawati A, Mujiyo M, Nugroho BDEP, Istiqomah NM, Irmawati V, Hasanah K, et al. Drought potential index using Normalized Difference Drought Index (NDDI) method based on GIS in Slogohimo, Wonogiri, Indonesia. IOP Conf Ser Earth Environ Sci. 2024;1314(1):012040.
8. Mu Q, Zhao M, Kimball JS, McDowell NG, Running SW. A remotely sensed global terrestrial drought severity index. Bull Am Meteorol Soc. 2013;94(1):83-98.
9. Stancalie G, Serban F. Agricultural drought monitoring using satellite-based products in Romania. In: Third International Conference on Telecommunications and Remote Sensing. 2014;1:100-106.
10. Tsakiris G, Loukas A, Pangalou D, Vangelis H, Tigkas D, Rossi G, Cancelliere A. Drought characterization. Drought Management Guidelines Technical Annex. 2007;58:85-102.