📄 earth_science

Attributing Runoff Variability to Climate Change and Land Use–Land Cover Change in the Hasdeo River Catchment Using SWAT

This study utilizes the SWAT model to demonstrate that significant land-use changes and climate variability have substantially increased runoff in the Hasdeo River catchment, with future projections under RCP scenarios indicating intensified seasonal streamflow and heightened flood risks.

Original authors: Anuradha Sahu, MK Nema, Susanta Das, VK Chandola

Published 2026-06-29

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Anuradha Sahu, MK Nema, Susanta Das, VK Chandola

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the Hasdeo River catchment as a giant, natural sponge sitting in the heart of India. This sponge is made of soil, forests, and fields, and its job is to soak up rain and release it slowly into the river. This study is like a detective investigation into what happens when you squeeze that sponge too hard or change its texture.

Here is the story of the paper, broken down into simple parts:

1. The Setting: A Sponge Under Pressure

The Hasdeo River area is a massive "sponge" (about 9,920 square kilometers). For a long time, this sponge was mostly covered in thick forests and open scrubland, which act like a soft, fluffy top layer that helps water soak in gently.

However, in recent decades, humans have been busy changing the sponge's surface. They've:

Cut down trees (deforestation).
Built more houses and cities (urbanization).
Dug up coal (mining).
Expanded farms.

Think of this like replacing the soft, fluffy top of a sponge with a hard, plastic sheet. When rain hits a plastic sheet, it doesn't soak in; it runs off immediately, rushing downhill.

2. The Detective Tool: The "SWAT" Model

To figure out exactly how much water is rushing off this changed landscape, the researchers used a computer program called SWAT (Soil and Water Assessment Tool).

Think of SWAT as a super-accurate video game simulator. The researchers fed the computer real data about the land, the soil, and the weather from 1991 to 2017. The computer then "played" the game to see how much water would flow through the river under those conditions.

The Result: The simulator was very good at the job. When they compared the computer's predictions to real-life measurements, they matched up very closely (about 80% accuracy). This gave the researchers confidence that their "game" was playing out reality correctly.

3. What Happened in the Past? (1991–2017)

The researchers looked at the "before and after" photos of the land.

The Change: Between 1991 and 2017, the area covered by buildings and farms grew significantly (by about 5% and 4% respectively). Meanwhile, the forests shrank (by about 5% to 6%).
The Effect: Because the "sponge" lost its fluffy forest cover and gained hard surfaces, the water had nowhere to go but straight into the river.
The Numbers: The study found that the amount of water running off the land increased by 146% in the computer simulation and 137% in real observations. Essentially, the river started carrying much more water than it used to, mostly because the land couldn't hold it anymore.

4. Looking into the Crystal Ball: Future Predictions

The researchers didn't just look at the past; they used a "crystal ball" (a climate model called HadGEM2) to guess what the future holds. They looked at three different scenarios for how the world might change by 2050:

Scenario A (RCP 2.6): A "gentle" future with low pollution.
Scenario B (RCP 4.5): A "moderate" future.
Scenario C (RCP 8.5): A "high" future with lots of pollution and heat.

The Prediction:
No matter which scenario happens, the river is expected to get wetter, especially during the rainy season.

Under the "gentle" scenario, the river flow could increase by 7%.
Under the "moderate" scenario, it could go up by 10%.
Under the "high" scenario, it could jump by 15%.

Even worse, the rainy season itself is predicted to get much heavier. The study predicts the rainy months could see 29% to 55% more water flowing than before. This is like turning a steady shower into a firehose.

5. Who is to Blame? (Climate vs. Land Use)

The researchers wanted to know: Is the extra water coming because it's raining more (Climate Change) or because we changed the land (Land Use)?

They found that it's a team effort, but the mix changes depending on how bad the future gets:

In a "gentle" future: The weather (rain) is the main boss, causing about 57% of the change.
In a "moderate" future: The interaction between the weather and the changed land becomes the biggest factor (about 59%).
In a "high" future: The combination of bad weather and our changed land (mining, cutting trees) creates a "perfect storm," accounting for nearly 70% of the change.

6. The Bottom Line

The study concludes that the Hasdeo River is becoming a faster, more powerful, and more unpredictable river.

The Risk: Because the water is rushing faster and in larger volumes, the risk of flooding is going up, especially in the lower parts of the river where the water collects.
The Warning: If we keep cutting down forests and expanding mining and cities without planning for this, the river might overflow its banks more often, threatening the people and farms living nearby.

In short: The land has become less like a sponge and more like a slide. When the rain comes, the water slides off too fast, and future storms are predicted to make that slide even steeper. The paper suggests we need to manage the land better to slow the water down again.

Technical Summary: Attributing Runoff Variability to Climate Change and Land Use–Land Cover Change in the Hasdeo River Catchment Using SWAT

Problem Statement
The Hasdeo River catchment (9,920 km²), a critical sub-basin of the Mahanadi River in Chhattisgarh, India, is undergoing severe environmental stress driven by anthropogenic activities. Extensive coal mining, rapid urbanization, and agricultural expansion have led to significant deforestation and landscape alteration. These Land Use–Land Cover (LULC) transformations, coupled with climate variability, are altering the hydrological cycle, specifically increasing surface runoff and reducing soil retention. This poses risks of abnormal flooding, reduced storage capacity in the Hasdeo Dam, and ecological imbalance. There is a need to quantify the relative impacts of LULC changes versus climate change on streamflow to inform water resource planning and catchment management.

Methodology
The study employed the Soil and Water Assessment Tool (SWAT), a semi-distributed hydrological model, to simulate runoff and attribute variability to specific drivers.

Study Area & Data: The model was applied to the Hasdeo catchment, divided into 23 sub-watersheds. Input data included Digital Elevation Models (DEM), soil maps (FAO), and daily meteorological data (1991–2017) from the India Meteorological Department (IMD). Hydrological data (daily streamflow) were sourced from the Central Water Commission.
LULC Analysis: Land use maps for 1991 and 2017 were generated using supervised classification of Landsat imagery (Landsat 5 and 8) in ArcGIS. Classification accuracy was validated using ground-truthing, yielding overall accuracies of 85% (1991) and 89% (2017) with Kappa coefficients of 0.81 and 0.86, respectively.
Model Calibration and Validation: The model was calibrated and validated using the SWAT-CUP interface with the SUFI-2 algorithm. The calibration period covered 1993–2010, and validation covered 2011–2017. Sensitivity analysis identified SOL_AWC (soil available water content) as the most sensitive parameter. Performance was evaluated using the coefficient of determination ( $R^2$ ), Nash–Sutcliffe Efficiency (NSE), and Kling–Gupta Efficiency (KGE).
Future Projections: Future climate scenarios were generated using the HadGEM2 Global Climate Model (GCM) under three Representative Concentration Pathways (RCP 2.6, 4.5, and 8.5) for the period 1991–2050. These climate projections were integrated with projected LULC changes to simulate future streamflow patterns.

Key Results

LULC Changes (1991–2017): The analysis revealed a 4.96% increase in built-up areas and a 3.83% increase in agricultural land. Conversely, open forest decreased by 6.70% and dense forest by 5.13%. Barren land and water bodies increased by 1.66% and 1.39%, respectively.
Model Performance: The SWAT model demonstrated good performance, with $R^2$ values of 0.84 (calibration) and 0.86 (validation), and NSE values of 0.73 and 0.71, respectively.
Historical Runoff Trends: Between 1991 and 2017, simulated runoff increased by 146.53%, while observed runoff increased by 137.79%. Seasonal average discharge rose from 257.8 m³/s (1991–2000) to 415.52 m³/s (2011–2017).
Spatio-Temporal Variability: Spatial analysis identified downstream sub-watersheds (specifically 12, 16, 20, 22, and 23) as having the highest streamflow, exceeding 1500 m³/s. Temporal analysis indicated a 10–15% increase in the frequency of extreme runoff events (3000–4000 m³/s class) after 2005 compared to the earlier period.
Future Streamflow Projections: Under future RCP scenarios, peak streamflow is projected to increase by 7% (RCP 2.6), 10% (RCP 4.5), and 15% (RCP 8.5) compared to historical levels. The wet season is expected to become significantly wetter, with increases of 29%, 43%, and 55% under the respective scenarios.
Attribution of Runoff Drivers: The study quantified the contributions of climate variability, LULC change, and their interaction:
- RCP 2.6: Climate variability was the dominant driver (57.57%), with LULC contributing 4.41% and interaction 38.02%.
- RCP 4.5: The interaction effect became dominant (59.42%), with climate at 33.71% and LULC at 6.87%.
- RCP 8.5: The interaction effect contributed 69.77% to runoff change. Direct LULC influence rose to 18.75%, while direct climate influence dropped to 11.48%.

Significance and Claims
The authors claim that this study provides a comprehensive evaluation of how climate variability and LULC changes jointly affect streamflow dynamics in a mining-intensive catchment. The findings confirm that while historical runoff increases were driven by both factors, future hydrological responses will be increasingly dominated by the interaction between climate forcing and land surface changes.

The paper asserts that the projected intensification of seasonal runoff and streamflow variability suggests a higher risk of flooding and altered water availability, particularly in downstream sub-watersheds. These findings are presented as valuable for policymakers and researchers involved in water resource planning, flood mitigation, and the regulation of mining activities. The study emphasizes the necessity of integrated watershed management strategies that account for both land-use planning and climate adaptation to mitigate future hydrological risks. The authors note limitations regarding the spatial resolution of meteorological data and the exclusion of explicit groundwater-surface water interactions, suggesting these as areas for future research to improve predictive accuracy.