India has experienced significant Land-Use and Land-Cover Change (LULCC) over the past few decades. In this context, careful observation and mapping of LULCC using satellite data of high to medium spatial resolution is crucial for understanding the long-term usage patterns of natural resources and facilitating sustainable management to plan, monitor and evaluate development. The present study utilizes the satellite images to generate national level LULC maps at decadal intervals for 1985, 1995 and 2005 using onscreen visual interpretation techniques with minimum mapping unit of 2.5 hectares. These maps follow the classification scheme of the International Geosphere Biosphere Programme (IGBP) to ensure compatibility with other global/regional LULC datasets for comparison and integration. Our LULC maps with more than 90% overall accuracy highlight the changes prominent at regional level, i.e., loss of forest cover in central and northeast India, increase of cropland area in Western India, growth of peri-urban area, and relative increase in plantations. We also found spatial correlation between the cropping area and precipitation, which in turn confirms the monsoon dependent agriculture system in the country. On comparison with the existing global LULC products (GlobCover and MODIS), it can be concluded that our dataset has captured the maximum cumulative patch diversity frequency indicating the detailed representation that can be attributed to the on-screen visual interpretation technique. Comparisons with global LULC products (GlobCover and MODIS) show that our dataset captures maximum landscape diversity, which is partly attributable to the on-screen visual interpretation techniques. We advocate the utility of this database for national and regional studies on land dynamics and climate change research. The database would be updated to 2015 as a continuing effort of this study.

Machine learning classifiers are being increasingly used nowadays for Land Use and Land Cover (LULC) mapping from remote sensing images. However, arriving at the right choice of classifier requires understanding the main factors influencing their performance. The present study investigated firstly the effect of training sampling design on the classification results obtained by Random Forest (RF) classifier and, secondly, it compared its performance with other machine learning classifiers for LULC mapping using multi-temporal satellite remote sensing data and the Google Earth Engine (GEE) platform. We evaluated the impact of three sampling methods, namely Stratified Equal Random Sampling (SRS(Eq)), Stratified Proportional Random Sampling (SRS(Prop)), and Stratified Systematic Sampling (SSS) upon the classification results obtained by the RF trained LULC model. Our results showed that the SRS(Prop) method favors major classes while achieving good overall accuracy. The SRS(Eq) method provides good class-level accuracies, even for minority classes, whereas the SSS method performs well for areas with large intra-class variability. Toward evaluating the performance of machine learning classifiers, RF outperformed Classification and Regression Trees (CART), Support Vector Machine (SVM), and Relevance Vector Machine (RVM) with a >95% confidence level. The performance of CART and SVM classifiers were found to be similar. RVM achieved good classification results with a limited number of training samples.

A high concentration of fluoride (F⁻) in drinking water is harmful and is a serious concern worldwide due to its toxicity and accumulation in the human body. There are various sources of fluoride (F⁻) and divergent pathways to enter into groundwater sources. High F⁻ incidence in groundwater was reported in Raigarh district of Central India in a sedimentary (Gondwana) aquifer system. The present study investigates the hydrogeochemistry of groundwater in the Tamnar area of Raigarh district to understand the plausible cause(s) of high F⁻ concentration, especially the source(s) and underlying geochemical processes. Groundwater samples, representing pre-monsoon (N = 83), monsoon (N = 20), and post-monsoon (N = 81) seasons, and rock samples (N = 4) were collected and analyzed. The study revealed that (i) groundwater with high F⁻ concentration occurs in the Barakar Formation, which has a litho-assemblage of feldspathic sandstones, shales, and coal, (ii) high F⁻ concentration is mainly associated with Na-Ca-HCO₃, Na-Ca-Mg-HCO₃, and Na-Mg-Ca-HCO₃ types of groundwater, (iii) the F⁻ concentration increases as the ratio of Na⁺ and Ca²⁺ increases (Na⁺: Ca²⁺, concentration in meq/l), (iv) F⁻ has significant positive correlation with Na⁺ and SiO₂, and significant negative correlation with Ca²⁺, Mg²⁺, HCO₃⁻, and TH, and (v) high F⁻ concentration in groundwater is found in deeper wells. Micas and clay minerals, occurring in the feldspathic sandstones and intercalated shale/clay/coal beds, possibly form an additional source for releasing F⁻ in groundwater. Feldspar dissolution coupled with anion (OH⁻ or F⁻) and cation (Ca²⁺ for Na⁺) exchange are probably the dominant geochemical processes taking place in the study area. The higher residence time and temperature of groundwater in deeper aquifers also play a role in enhancing the dissolution of fluorine-bearing minerals. Systematic hydrogeochemical investigations are recommended in the surrounding area having a similar geologic setting in view of the potential health risk to a large population.