Integrating Remote Sensing and GIS for LULC Studies in the Warana River Basin

Land is a vital natural resource that serves as the foundation for human life, agriculture, and economic development. It provides space and raw materials for various human activities, but increasing pressure from population growth, urbanization, and intensive use has led to continuous changes in land use patterns. Proper land utilization depends on the ecological and geological suitability of an area, and its sustainable management is crucial for maintaining environmental balance and food security (Twisa & Buchroithner, 2019). Land Use and Land Cover (LULC) change is a significant indicator of environmental transformation influenced by both natural and human factors, including population growth, topography, soil type, and climate (Tewabe & Fentahun, 2020). These changes affect biodiversity, water cycles, and land productivity. LULC analysis helps assess the impacts of activities like agriculture, deforestation, and urbanization on natural ecosystems (Seyam et al., 2023). Change detection techniques allow for monitoring landscape alterations and human-nature interactions to support better decision-making (Rawat & Kumar, 2015). Remote Sensing (RS) and Geographic Information Systems (GIS) are key tools in studying LULC. They allow for accurate mapping, analysis, and monitoring of spatial and temporal changes in land use, offering a cost-effective and reliable method for environmental management (Herold et al., 2003; Jensen, 1996). Satellite imagery provides multi-spectral, geo-referenced data to detect changes in land cover, supporting studies on urban expansion, agriculture, and industrial growth (Sekela & Manfred, 2019). The Warana River Basin, located in Kolhapur and Sangli districts of Maharashtra, is a significant geographical feature for this study. The river originates in the Sahyadri ranges and flows southeast for about 129 km before joining the Krishna River at Haripur. Its watershed is important for regional agriculture and land use planning (Bist et al., 2021). Understanding LULC changes in such river basins is crucial for water resource management, biodiversity conservation, and climate adaptation strategies.

MATERIAL & METHODS

1. Study Area

1.1 Location

The Warana River is located in the Western Ghats of Maharashtra, India. It flows through the Sangli district, originating near Prachitigad Fort in the Sahyadri range. The river ultimately merges with the Krishna River. The study area lies between 16º0'47'' N to 17º0'15'' N latitude and 73º30'15'' E to 74º0'30'' E longitude. The Warana River basin is located in a tropical and subtropical climatic zone, hence it has pleasant temperatures all year. The temperature in the basin can fluctuate between 30 °C and 40 °C. The Warana River basin has an area 1965 square km.

1.2 Climate and Topography

The Warana River Basin in Western Maharashtra experiences a warm and humid climate with three distinct seasons: summer, monsoon, and winter. Summers (March to May) are hot and humid, with daytime temperatures ranging between 30°C to 40°C and nighttime temperatures between 20°C to 25°C. The monsoon season (June to September) brings significant rainfall, with average annual precipitation of approximately 1,440 mm, accompanied by milder temperatures ranging from 25°C to 30°C during the day and around 20°C at night. Winters (November to February) are cooler and dry, with daytime temperatures ranging from 10°C to 25°C and nighttime lows between 5°C and 15°C, depending on the location. Topographically, the Warana Basin is characterized by varied elevation, ranging from 455 meters to over 1,032 meters above sea level. The region includes hilly terrain and undulating landscapes, with the Amba Ghat being the highest point at approximately 1,150 meters, located near the border of Kolhapur and Ratnagiri districts. This elevation variation influences the basin’s microclimates and water flow patterns, playing a significant role in the area's land use and agricultural practices.

2. Data Required

2.1 Precipitation

 Data on rainfall patterns and amounts over time from CHIRPS (Climate Hazards Group   Infrared Precipitation with Station data) for the last 30 years. The region receives significant rainfall from the southwest monsoon, contributing to river flow. The average annual rainfall is 600-2500mm, with a potential risk of flooding during heavy monsoon periods.

2.2 Soil

Data on soil properties such as texture from Open Land Map, a free and open-source global soil information system. The soil along the Warana River ranges from shallow to moderately deep, with loamy to clayey textures. Black soils in the plains are rich in minerals, while red lateritic soils dominate the hilly regions. Key soil parameters include pH, organic matter, and water retention capacity. Open Land Map, a free and open-source global soil information system, provides data on soil qualities such as texture.  Soil map helps to guide decisions on agricultural practices by identifying the best soil for specific crops, assist in urban planning by determining areas suitable for construction, and support environmental conservation efforts by identifying regions vulnerable to erosion or degradation. Additionally, soil maps play a role in forestry, climate change modeling, and land suitability analysis, ensuring that land development is sustainable and aligned with the natural characteristics of the soil.

2.3 Land Use and Land Cover (LULC)

 Data on the physical and biological characteristics of the Earth's surface by supervised classification of Landsat 8 Images (from USGS Earth Explorer)

2.4 Drainage System

The Warana River is part of the Krishna River basin, with a high drainage density in the hilly areas. The river displays meandering patterns and alluvial deposits in its lower reaches.

2.5 Landsat-8

Landsat 8 is part of the Landsat program, which has been providing valuable Earth observation data since 1972. Landsat 8 was launched on February 11, 2013, by NASA and the U.S (NASA Landsat Science). Geological Survey (USGS). It is equipped with two key instruments: the Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS). These instruments capture high-resolution images of Earth's surface, with applications in a wide range of fields, including agriculture, forestry, water resources, land-use planning, and disaster management.

3. Softwares Used

3.1. ArcGIS

 ArcGIS is a comprehensive platform for working with maps and geographic information. It allows users to create, manage, analyze, and visualize spatial data. It supports vector and raster data formats, and offers extensive tools for spatial analysis, geostatistics, network analysis, and data management.

3.2 Google Earth Pro

Google Earth Pro is a powerful desktop application developed by Google that provides detailed satellite imagery, maps, terrain data, and 3D representations of the Earth. It allows users to explore geographical features, measure distances, create custom maps, and view historical imagery to analyze changes over time. This user-friendly resource is often a useful intermediary for learners who are interested in learning more about GIS and want to start with more basic processes and tools. Google Earth Pro can also be leveraged to view its extremely high-resolution satellite imagery, upload or download geospatial data in its native interoperable file format (KML), and also find locations (e.g. for simple geocoding).

4. Development of Land Use Land Cover (Lulc) Maps

In recent decades, remote sensing has become a crucial tool for Land Use and Land Cover (LULC) change detection due to its consistent image quality and frequent revisit capabilities. Satellite imagery offers a reliable and cost-effective source of LULC data, enabling researchers to monitor changes over time. High-resolution LANDSAT satellite data from the USGS platform was used to cover the study area across multiple spectral bands, ensuring sufficient spectral, spatial, and temporal resolution for accurate analysis. For this study, the supervised classification technique was employed to categorize land into five classes: agriculture, habitation, waterbodies, barren land, and forest areas—based on the Anderson LULC classification system. LULC maps of the Warana River Basin were generated for the years 2014 and 2024, and changes over the 10-year period were examined and quantified.

4.1 Supervised Classification

Supervised classification in ArcGIS is a process of categorizing pixel data in imagery based on user-defined training samples. This technique is widely used in remote sensing and GIS to classify land cover, vegetation types, urban areas, and other features in satellite or aerial images. It involves creating "training samples" where you assign known classes to certain areas in an image (like water, forest, agriculture), and then allowing the software to apply these classifications to the rest of the image based on spectral similarity.  ArcGIS provides several tools within its Image Analyst and Spatial Analyst extensions that support supervised classification, including: Training Sample Manager Tool, Conversion Tool, Raster Tool, Vector Machine, Accuracy Assessment Tool. ArcGIS's tools for supervised classification enable detailed and accurate land cover analysis, benefiting applications like urban planning, environmental monitoring, and agricultural assessment.

5. Accuracy assessment

Accuracy assessment is a critical step in validating land use and land cover (LULC) classification maps created using remote sensing. In ArcGIS, it involves comparing the classified map with ground-truth data to measure how accurately land cover types have been identified. Key statistical measures used include overall accuracy, user’s accuracy, producer’s accuracy, and the Kappa coefficient, all of which help determine the reliability of the classification process. The Kappa coefficient (κ) is a widely used statistical metric that measures the degree of agreement between the classified data and reference data, beyond what would occur by chance. It is calculated using the formula κ = (po – pe) / (1 – pe), where po is the observed agreement and pe is the expected agreement by chance. Values of Kappa range from -1 to 1, with a value close to 1 indicating strong agreement. Based on Islami (2022), a Kappa value between 0.81–1.00 signifies almost perfect agreement, while values below 0.20 indicate slight agreement. This makes Kappa particularly useful for assessing the true accuracy of classification efforts. In addition to Kappa, overall accuracy represents the proportion of correctly classified pixels out of the total sample, providing a simple but effective summary of classification performance. User’s accuracy reflects how often pixels classified into a given category actually represent that category on the ground, while producer’s accuracy indicates how well reference land cover classes are represented on the map. These metrics are typically derived from an error/confusion matrix, which compares classified data against known ground truth points. To validate classified outputs, ground truthing is essential. It involves field visits and data collection—including photographs, GPS coordinates, and physical observations—to confirm land cover types like agriculture, forest, urban, or water bodies. Various sampling methods such as random, systematic, or stratified random sampling are used to gather accurate reference data. Ground truthing helps reduce classification errors, improves model training, and ensures the classified maps reflect real-world land dynamics accurately.

RESULT AND DISCUSSION

The comparative analysis of Land Use Land Cover (LULC) changes using Landsat 8 satellite imagery between 2014 and 2024 reveals profound transformations in land use patterns driven by socio-economic and environmental factors. The study indicates a significant expansion of urban areas, particularly in regions with rapid population growth and economic activities. This urban growth has often occurred at the expense of agricultural land and natural vegetation, leading to habitat fragmentation and biodiversity loss.

Figure. LULC Map for Year 2014

Figure. LULC Map for Year 2024

Area of Water Bodies decreased from 29.95 sq.km (1.52%) in 2014 to 25.40 sq.km (1.29%) in 2024. The reduction in water bodies can be attributed to urban encroachment, over-extraction of water resources, and reduced rainfall in the region. Climate change and increased evaporation rates may also have contributed to this decline. Area of Agriculture is decreased from 650.80 sq.km (33.12%) in 2014 to 460.12 sq.km (23.42%) in 2024. Agricultural land shrank due to urban expansion and reforestation efforts. Fertile lands were often converted into forest zones and urban settlements, reflecting shifts in land-use priorities driven by developmental and conservation policies. Area of Habitation is increased from 85.14 sq.km (4.33%) in 2014 to 151.61 sq.km (7.72%) in 2024.The rise in habitation area is due to rapid urbanization and infrastructure development. Population growth and industrialization likely played a key role in expanding urban settlements. Area of Forest increased from 1198.77 sq.km (61.01%) in 2014 to 1325.51 sq.km (67.46%) in 2024. Conservation and reforestation initiatives have driven the increase in forest cover. Policies promoting afforestation, biodiversity conservation, and restoration of degraded land contributed to this growth.  Area of Barren Land increased from 0.11 sq.km (0.01%) in 2014 to 2.13 sq. km (0.11%) in 2024. The growth in barren land is linked to mining activities, urban expansion, and improper land conservation practices. The conversion of other land types into unproductive land due to human activities has exacerbated this trend. The observed changes highlight the interplay of human activities and environmental management. Urban growth and development have reduced agricultural and water resources, while targeted conservation efforts have improved forest cover. Mining and infrastructure projects have also contributed to the rise in barren land. These trends underscore the need for sustainable planning to balance development with environmental preservation.

Figure. Status of LULC in 2014

Figure. Status of LULC in 2024