Dr. Swapan Kumar Masanta

 

Brief Details:

Dr. Swapan Kumar Masanta I have completed my Bachelor of Technology (B.Tech) in Agricultural Engineering at Bidhan Chandra Krishi Viswavidyalaya, Mohanpur, Nadia, West Bengal, where I obtained gold medal for highest CGPA. I pursued a Master of Engineering (M.E) in Water Resources and Environmental Engineering at the Indian Institute of Science, Bangalore. Subsequently, I embarked on my Doctor of Philosophy (Ph.D) journey, specializing in Water Resources and Environmental Engineering at the same esteemed institution, the Indian Institute of Science, Bangalore. During my PhD, I published two papers in International Journal of Climatology (IJOC) and one paper in Journal of Hydrology (JH). Currently, I am working as consultant of R&D in Xceedance. Where, I am preparing probabilistic map for flood, wildfire, cyclone and drought considering climate change scenarios for use as hazard risk data in insurance industry.

My research interests are, the prediction of hydrological processes and extremes in climate change scenarios; analysis of risk associated with hydrological extremes; regionalization for estimating hydrological variables in ungauged locations.

Research

The research presentation will cover the regional to basin scale research during my Ph.D., the technology developed and transferred to the grassroots level, the postdoctoral research at the USA, the research outcomes in terms of publication and fund acquisitions, my capacity building activities, academic and industrial collaborations, my short and long-term goals, and significance of all these experiences for WRD&M department, IIT Roorkee. In the regional scale analysis, I put an effort into investigating the spatiotemporal variability of the groundwater level and determined the dominant hydrogeological and climatic controls regulate the observed variability. Then I jumped into the basin-scale study with the takeaways from the regional-scale study to go for a concurrent groundwater exploration and recharge potential mapping study taking the secondary data available. Simultaneously, I worked on field and laboratory experiments to generate my primary dataset, which helped in conducting the physical and machine learning (ML)-based modeling studies. In that process, I developed a generalizable pedotransfer function for the estimation of saturated hydraulic conductivity using the ML algorithm. Subsequently, developed the physical-based models for recharge and groundwater head and bridged that to the Bayesian Decision Network to develop a mobile app-based decision support system for groundwater development and management. The technology developed got two funding for technology transfer to address water issues in Maharashtra and Odisha. The postdoctoral research addresses the nitrate concentration issues in drinking water wells through hydrogeological analysis under the edges of Forensic Hydrology. These research activities have generated quite a few academic and industrial collaborations and some high-impact international publications along with a few publications and a patent/ copyright in the pipeline. Additionally, these research activities have created a huge scope of short and long-term research, which are in line with the major mandates of the WRD&M Department and may create an impactful research infrastructure.

Teaching

The analysis of hydrological change is vital for both science and practical applications. Long-term hydrological data offer profound insights into dynamic water systems. One key objective is trend detection, crucial for re-evaluating infrastructure design when the assumption of stationary hydrology proves incorrect, preventing inadequate systems or excessive costs. Understanding hydrological change is also essential for assessing human-induced environmental impacts like urbanization, deforestation, greenhouse gas emissions, and dam construction, ensuring harmonious coexistence with nature. Moreover, studying shifts in hydrological extremes, such as floods and droughts, aids in anticipating their frequency and intensity, enhancing management and adaptation strategies. In sum, analyzing hydrological change is pivotal for resilience, environmental stewardship, and addressing water-related challenges.

 

Various approaches can be employed to detect trends and non-stationarity in hydrological data, each with its own considerations. Parametric testing, rooted in classical statistics, assumes an underlying data distribution (often normal) and independence between data points. However, hydrological data rarely adhere to these assumptions. Transforming data and analyzing annual series may be necessary for parametric techniques, limiting their applicability. Non-parametric and distribution-free methods impose fewer assumptions about data distribution but still rely on independence assumptions, making them suitable for daily or hourly series. Given the unpredictability of variability, employing multiple tests is a prudent approach to comprehensive change detection.

 

In the context of trend detection, two statistical approaches are mostly employed. The first approach utilizes slope-based tests, specifically least squares linear regression and Sen's robust slope estimator to identify trends. The second approach employs rank-based tests, namely Mann-Kendall and Spearman rank correlation for trend detection. These methods provide valuable tools for analyzing and identifying trends within datasets. However, some necessary modifications are required to consider the effect of serial correlation, including pre-whitening, block resampling techniques such as block bootstrap with Mann-Kendall test.

 

The discussion involves descriptions of three statistical tests for trend detection in hydrologic and climatic data. The first is the Least Squares Linear Regression test, a parametric method sensitive to data normality and used to identify linear trends. The Mann-Kendall test, a non-parametric rank-based test, examines monotonic trends by assessing data sequences. Finally, Sen's Slope method calculates slope estimates to determine increasing or decreasing trends based on the median of these estimates. Each of these tests is demonstrated with an example with hydrological time series data.

Google Scholar Link: 

 https://scholar.google.com/citations?user=0Y8oe-cAAAAJ&hl=en