Drought Analysis Using Google Earth Engine: Rainfall and NDVI Anomaly Mapping in Kerala

Akshitha K J
17 hours ago
10 min read

Abstract

Drought is one of the most complex and spatially variable climatic hazards, affecting hydrological systems, agricultural productivity, and ecological stability across different climatic regions. The increasing variability of rainfall patterns and the growing dependence on climate-sensitive sectors have intensified the need for accurate and spatially explicit drought monitoring methods. Remote sensing and geospatial technologies provide valuable tools for assessing drought conditions by enabling continuous monitoring of climatic and vegetation parameters over large areas.

This study presents a comprehensive drought assessment for the year 2023 using satellite derived rainfall and vegetation indicators in Palakkad. Annual rainfall was estimated using the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS), and long term rainfall climatology (2001–2020) was computed to represent baseline conditions. Vegetation response was evaluated using the Normalized Difference Vegetation Index (NDVI) derived from MODIS data. Rainfall anomaly and NDVI anomaly were calculated to quantify deviations from normal climatic and ecological conditions. Both anomaly layers were classified into five categories representing varying levels of deficit and stress. A composite drought severity index was developed by integrating rainfall and vegetation indicators to provide a holistic representation of drought conditions.

The study demonstrates the effectiveness of satellite-based datasets in capturing spatial variability in drought patterns and highlights the importance of integrating climatic and vegetation indicators for comprehensive drought monitoring. The methodology adopted in this study can serve as a framework for regional drought assessment and can support planning strategies aimed at climate resilience and sustainable resource management.

Introduction

Drought is a natural climatic phenomenon characterized by a prolonged deficiency of precipitation relative to the long-term average, resulting in water scarcity and environmental stress. Unlike other natural disasters that occur abruptly, drought develops gradually and can persist over extended periods, making it difficult to detect and manage effectively. Its impacts are multidimensional, influencing agriculture, water resources, ecosystems, and socio-economic systems. In regions where livelihoods depend heavily on rainfall, drought can lead to crop failure, reduced water availability, and ecological degradation.

Drought can be broadly categorized into meteorological, agricultural, hydrological, and socio-economic drought, depending on the nature of impacts and indicators used for assessment. Meteorological drought refers to a deficit in precipitation compared to normal levels, while agricultural drought relates to soil moisture deficiency affecting crop growth. Hydrological drought occurs when reduced precipitation leads to lower streamflow and groundwater levels, and socio-economic drought arises when water scarcity affects human activities and economic systems.

Traditional drought monitoring methods rely primarily on meteorological station data. While these observations provide accurate point measurements, their sparse spatial distribution often limits their ability to represent regional variability. Remote sensing techniques overcome this limitation by providing consistent, spatially continuous data, enabling monitoring of environmental variables such as rainfall, soil moisture, and vegetation condition.

Satellite-derived rainfall datasets have significantly improved drought monitoring by providing near-real-time information on precipitation patterns. Similarly, vegetation indices such as the Normalized Difference Vegetation Index (NDVI) have been widely used to monitor vegetation health and assess ecological response to climatic stress. NDVI is sensitive to changes in chlorophyll content and canopy structure, making it a reliable indicator of plant vigor and moisture availability.

The integration of rainfall anomalies and vegetation anomalies provides a comprehensive approach to drought assessment, as it captures both the climatic drivers and the ecological impacts of drought. Such integrated approaches are particularly important in monsoon dominated regions where rainfall variability directly influences agricultural productivity and ecosystem stability.

Palakkad district, located in a climatically sensitive region, frequently experiences fluctuations in rainfall patterns that affect agriculture and water resources. Assessing drought conditions in this region is therefore crucial for understanding climate variability and supporting resource management strategies. This study aims to evaluate drought conditions for the year 2023 using satellite-derived rainfall and NDVI data and to develop a composite drought severity map representing spatial drought patterns across the district.

Objectives

The main objective of this study is to assess drought conditions using rainfall and vegetation indicators and to map drought severity spatially. The specific objectives are:

To analyze the spatial distribution of annual rainfall for the year 2023.
To compute long-term mean rainfall and identify deviations from normal conditions.
To evaluate vegetation condition using NDVI and assess vegetation stress through anomaly analysis.
To classify rainfall and NDVI anomalies into drought-relevant categories.
To generate a composite drought severity map integrating climatic and vegetation indicators.
To demonstrate the applicability of geospatial techniques for regional drought monitoring.

Study Area

Palakkad district is located in the northeastern part of Kerala, India, and lies between the Western Ghats and the plains of Tamil Nadu. The district forms a unique geographical corridor that influences regional climate and weather patterns. The terrain of the district is characterized by a mix of mountainous regions, forested areas, river basins, and agricultural plains.

The climate of the region is tropical monsoon, with the majority of rainfall occurring during the southwest monsoon season from June to September, followed by the northeast monsoon from October to December. Rainfall variability plays a critical role in determining agricultural productivity, groundwater recharge, and ecosystem health in the district. Seasonal fluctuations in rainfall often result in periods of moisture deficit, making the region susceptible to drought conditions.

Agriculture is the primary economic activity in the district, with major crops including paddy, coconut, and pulses. The dependence of agriculture on rainfall makes the district particularly vulnerable to climatic variability. In addition, the district contains diverse ecosystems ranging from forested highlands to cultivated lowlands, which respond differently to moisture stress. The spatial heterogeneity of climate and land cover makes Palakkad an ideal case study for drought assessment using remote sensing and geospatial techniques.

Materials and Methodology

Materials

CHIRPS Rainfall Dataset

The Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) provides high-resolution precipitation estimates by combining satellite observations with groundbased station data. The dataset has a spatial resolution of approximately 5 km and is widely used for drought monitoring and hydrological studies.

MODIS NDVI Dataset

The MODIS MOD13Q1 product provides 16-day composite NDVI data at a spatial resolution of 250 m. NDVI is calculated from red and near-infrared reflectance and represents vegetation greenness and photosynthetic activity. It is widely used to assess vegetation health and detect stress caused by moisture deficiency.

Administrative Boundary Data

The GAUL administrative boundary dataset was used to extract the district boundary and clip all datasets to the study area.

Methodology

Rainfall Analysis

Daily rainfall data were obtained from the CHIRPS dataset and aggregated to compute total annual rainfall for the year 2023. To establish the climatological baseline, long-term mean annual rainfall was calculated using multi-year data for the period 2001–2020.

Rainfall anomaly was computed to quantify deviations from normal precipitation conditions. The anomaly represents the difference between observed rainfall and the long-term mean and was calculated using the following expression:

Rainfall Anomaly = 𝑅2023−𝑅𝐿𝑇

where :

𝑅2023= total annual rainfall for 2023

𝑅𝐿𝑇= long-term mean annual rainfall

Positive anomaly values indicate above-normal rainfall, while negative values represent rainfall deficit conditions. Thus, the rainfall anomaly map highlights the spatial variability of meteorological drought by detecting areas where precipitation deviates significantly from typical climatic patterns.

NDVI Analysis

Vegetation condition was assessed using the Normalized Difference Vegetation Index (NDVI) derived from MODIS imagery. NDVI is calculated based on the reflectance difference between the near-infrared (NIR) and red spectral bands and is expressed as:

NDVI =𝑁𝐼𝑅−𝑅𝑒𝑑/𝑁𝐼𝑅+𝑅𝑒𝑑

NDVI values range from −1 to +1, where higher values indicate dense and healthy vegetation, while lower values correspond to sparse vegetation, bare soil, or built-up areas.

Mean annual NDVI for 2023 was computed to represent vegetation conditions during the study year. Long-term mean NDVI was calculated using multi-year data (2001–2020) to establish baseline vegetation conditions.

NDVI anomaly was then derived using:

NDVI Anomaly = 𝑁𝐷𝑉𝐼2023−𝑁𝐷𝑉𝐼𝐿𝑇

where:

𝑁𝐷𝑉𝐼2023= mean NDVI for 2023 𝑁𝐷𝑉𝐼𝐿𝑇= long-term mean NDVI

The NDVI anomaly indicates vegetation response to climatic variability. Negative values represent vegetation stress or reduced greenness, while positive values indicate improved vegetation conditions relative to the long-term average.

Classification of Anomalies

To facilitate spatial interpretation, rainfall anomaly and NDVI anomaly were classified into five categories representing increasing levels of rainfall deficit and vegetation stress. Classification helps standardize the interpretation of continuous data and enables comparison between climatic and ecological indicators.

The classification scheme highlights:
Near-normal conditions
Mild deficit
Moderate deficit
Severe deficit
Extreme deficit

This step converts quantitative anomaly values into meaningful thematic classes for mapping and analysis.

Drought Severity Mapping

A composite drought severity index was generated by integrating the classified rainfall anomaly and NDVI anomaly layers. The integration approach captures both meteorological and ecological aspects of drought, ensuring a more comprehensive assessment than single-parameter analysis.

The combined severity index represents overall drought intensity by identifying areas where rainfall deficit coincides with vegetation stress. Higher index values indicate greater drought severity, while lower values represent milder drought conditions.

This integrated mapping approach enables the identification of spatial drought patterns and supports the evaluation of regional vulnerability across the study area.

To improve spatial continuity and visual representation of rainfall distribution, IDW Interpolation was applied to rainfall and long term rainfall surfaces. Rainfall Anomaly and Severity maps were retained in their original raster format to preserve calculated index values.

Results and Discussion

Rainfall Distribution in 2023

The spatial distribution of annual rainfall for 2023 indicates that the district received precipitation ranging from approximately 949 mm to 2526 mm, with a mean rainfall of about 1738 mm. Higher rainfall zones are predominantly observed along the windward and elevated regions, while comparatively lower rainfall pockets are visible in interior areas. The spatial variability reflects localized climatic influences and topographic controls on precipitation distribution.

Long-Term Mean Rainfall (2001–2020)

The long-term mean rainfall map represents the baseline climatic condition of the district. Rainfall during the reference period ranges between 1111 mm and 3011 mm, with an average of 2061 mm. The spatial pattern indicates that the region typically experiences high monsoonal precipitation, confirming its humid tropical climate characteristics.

Rainfall Anomaly

The rainfall anomaly analysis reveals a district-wide negative deviation from the long term average, with anomaly values ranging from approximately −29 mm to −911 mm and a mean anomaly of −378 mm. This indicates that 2023 experienced a significant rainfall deficit compared to normal climatic conditions. Areas exhibiting anomalies below −600 mm represent severe rainfall shortages, suggesting the occurrence of meteorological drought conditions.

NDVI Distribution

Figure 5: NDVI 2023 — **Figure 5:** NDVI 2023

The NDVI distribution map for 2023 shows values ranging from 0.03 to 0.87, with a mean of 0.66, indicating generally healthy vegetation cover across the district. Dense vegetation is predominantly observed in forested and agricultural zones, whereas low NDVI values correspond to built-up areas, barren lands, and water bodies.

NDVI Anomaly

The NDVI anomaly values range between −0.25 and 0.18, with an average of 0.025, indicating mixed vegetation response across the district. Negative anomaly zones highlight areas experiencing vegetation stress, likely influenced by rainfall deficit and soil moisture reduction, while positive anomalies suggest localized resilience or favourable micro-climatic conditions.

Integrated Drought Severity Map

Derived by combining rainfall anomaly and NDVI anomaly classifications, provides a comprehensive representation of drought conditions. Severity values range from 2 to 5, with a mean of 3.73, indicating that most areas fall under moderate to severe drought categories, while localized zones exhibit extreme drought conditions. This spatial pattern confirms the combined influence of meteorological deficit and vegetation stress across the district.

Conclusion

This study presented a comprehensive geospatial assessment of drought conditions in Palakkad for the year 2023 through the integration of rainfall and vegetation indicators derived from satellite datasets. By comparing annual rainfall with long-term climatic averages and evaluating vegetation response through NDVI anomalies, the study aimed to identify the spatial variability of drought intensity and understand its environmental implications.

The rainfall analysis clearly indicated that the district experienced a below-normal precipitation year, with rainfall values consistently lower than the long-term mean across most parts of the region. The rainfall anomaly map confirmed a widespread negative deviation, demonstrating that meteorological drought conditions prevailed during the study period. However, the spatial distribution of drought was not uniform, highlighting the importance of localized assessment rather than district-wide generalization.

The integrated drought severity index revealed a distinct spatial gradient across the district. The western zone exhibited extreme drought conditions, despite recording relatively higher absolute rainfall compared to other regions. This pattern reflects the significant departure from its normally high rainfall baseline, indicating that drought severity is governed more by deviation from climatic norms than by absolute precipitation amounts. The central region experienced severe drought conditions, where moderate rainfall deficits combined with observable vegetation stress resulted in increased drought intensity. In contrast, the eastern zone showed predominantly moderate drought severity, suggesting comparatively lower deviation from its typical rainfall regime and a relatively reduced ecological impact.

The comparison between rainfall anomaly and NDVI anomaly demonstrated a strong relationship between precipitation variability and vegetation response. Areas with larger rainfall deficits corresponded with zones of vegetation stress, confirming the sensitivity of vegetation health to moisture availability. At the same time, spatial differences in NDVI response highlighted the role of land cover characteristics, soil moisture retention, and local environmental conditions in moderating drought impacts.

Overall, the integrated approach adopted in this study effectively captured both the meteorological and ecological dimensions of drought, providing a more holistic understanding of drought dynamics than single-indicator assessments. The findings underscore the value of combining climatic and vegetation indicators for spatial drought monitoring, as this approach improves the identification of vulnerable zones and supports more targeted mitigation planning.

From a regional perspective, the study demonstrates that drought intensity in 2023 was characterized by spatial heterogeneity, with the highest impact concentrated in the western part of the district and progressively decreasing toward the east. This pattern emphasizes that drought vulnerability is shaped by the interaction of climatic variability, baseline environmental conditions, and land surface characteristics.

In conclusion, the research confirms that geospatial techniques provide an effective framework for drought assessment at the district scale. The integration of rainfall and NDVI anomalies offers a reliable method for identifying drought severity patterns and understanding their spatial drivers. The results highlight the need for localized drought management strategies, improved water resource planning, and continuous monitoring using remote sensing tools to enhance resilience against future climate variability.

REFERENCES

Thomas, J., Joseph, S., Thrivikramji, K. P., & Arunkumar, K. S. (2015). Rainfall trends and variability over Kerala, India. Theoretical and Applied Climatology, 120(3–4), 537– 548.

Sreekanth, P. D., & Nair, M. T. (2018). Drought assessment using rainfall anomaly and NDVI over Kerala. Journal of Earth System Science, 127(6), 1–12.

Raju, K. S., Kumar, D. N., & Mishra, A. (2014). Drought monitoring in India using standardized precipitation index and remote sensing data. International Journal of Climatology, 34(4), 1336–1346.

Pai, D. S., Sridhar, L., Rajeevan, M., Sreejith, O. P., Satbhai, N. S., & Mukhopadhyay, B.

(2014). Development of a new high spatial resolution rainfall data set for India. Meteorological Applications, 21(3), 555–567.

Kumar, V., Jain, S. K., & Singh, Y. (2010). Analysis of long-term rainfall trends in India. Hydrological Sciences Journal, 55(4), 484–496.

Jain, S. K., Keshri, R., Goswami, A., & Sarkar, A. (2010). Application of meteorological and vegetation indices for evaluation of drought impact. Hydrological Processes, 24(17), 2543–2552.

Murthy, C. S., Rao, P. V. N., & Raju, P. L. N. (2007). Agricultural drought assessment using NDVI and rainfall anomaly. International Journal of Remote Sensing, 28(15), 3677–3692.

Rao, B. B., Chowdary, P. S., & Sandeep, V. M. (2017). Spatiotemporal analysis of drought using remote sensing and GIS in India. Natural Hazards, 85(2), 1225–1243.

Mishra, A. K., & Singh, V. P. (2010). A review of drought concepts. Journal of Hydrology, 391(1–2), 202–216.

Prasad, A. K., Chai, L., Singh, R. P., & Kafatos, M. (2006). Crop yield estimation using remote sensing and surface parameters. International Journal of Applied Earth Observation and Geoinformation, 8(1), 26–33.

Rouse, J. W., Haas, R. H., Schell, J. A., & Deering, D. W. (1974). Monitoring vegetation systems in the Great Plains with ERTS. NASA Special Publication, 351, 309–317.