The Indian Institute of Remote Sensing is a premier institute for research, higher education and training in the field of Remote Sensing, Geoinformatics and GPS Technology for Natural Resources, Environmental and Disaster Management under the Indian Department of Space, which was established in the year 1966. It is located in the city of Dehradun, Uttarakhand.
What if a satellite could detect a water-logged surface beneath the dense forests, which is invisible to ordinary optical cameras? A new Radar Imaging Satellite, ‘NISAR,’ is making this possible, helping scientists and researchers to understand the forests, wetlands, vegetation, and environmental changes in a completely new way. NISAR is an exciting new Earth-observation satellite mission developed jointly by NASA and ISRO to better understand our changing planet. Unlike ordinary satellites that capture pictures using sunlight, NISAR uses advanced radar technology called Synthetic Aperture Radar (SAR), allowing it to observe Earth Day and night, even penetrating through clouds. With this technology, scientists can track subtle environmental changes, including shifting ground after earthquakes, changing forests, expanding wetlands, evolving agricultural landscapes, and glaciers.
Accurate prediction of extreme rainfall events (EREs), such as cloudbursts, over the North-West Himalayan region (NWH) remains challenging due to the high sensitivity of model simulations to the choice of physics parameterization schemes. This study attempts to find the most suitable Planetary Boundary Layer (PBL) scheme in simulating EREs (cloud burst events) over various locations in Uttarakhand and Himachal Pradesh based on a sequence of events happened during 19 to 20th August 2022. For the comprehensive sensitivity analysis, a series of model experiments have been performed corresponding to different PBL schemes but for identical other physics options and initial conditions using Weather Research and Forecasting model (WRF). Our analysis revealed that the YSU scheme showed relatively better performance than the other tested schemes with good correlation, comparatively low relative mean absolute error (rMAE) and moderate RMSE with respect to reference data over majority of the locations. Simulated rainfall in all schemes, including YSU show low correlation with reference data over Himachal Pradesh (r < 0.15 over Jhulara and Banet, r < 0.5 over Gohar (Mandi)). It is worth mentioning that PBLH simulations by YSU outperforms other schemes over those locations where YSU reproduce better simulations of rainfall. The findings pertaining to the performance of the PBL schemes are evocative of the specific synoptic event analysed here, which embedded multiple EREs that produced rain over various locations in the NWH region. Further, we examined the relationship between model simulated rainfall and PBLH during EREs and found that high rainfall ensues with a lag of around 7-10 h from PBLH drop over majority of the locations. The role of PBL physics in the formation of EREs is least explored in previous studies. Regional differences in skill scores noted in the present study and the lead-lag relationship between PBLH and rainfall may guide region-specific improvements in PBL parameterizations for improved predictions of EREs over topographically complex NWH region.
Accurate mapping of crop types is essential for addressing food security, crop inventory and supporting farmers in decision making to improve production and manage agricultural practices. Generally, crop mapping studies either ignore minority crops or aggregate them into 'other crops' class during classification. This study addresses the challenge of class imbalance in crop type classification using remote sensing imagery, focusing on agricultural fields in Jhansi District, Uttar Pradesh, India, during the Rabi cropping season. A one-dimensional convolution neural network (1D CNN) based sequential deep learning model was employed using multitemporal Sentinel 2 imagery. Algorithm level and data level balancing techniques were examined. Categorical cross-entropy, focal loss, and class weighted loss functions were compared to reduce class imbalance, with the latter proving most effective. Additionally, data level balancing methods including undersampling, oversampling, and hybrid approaches were investigated, with undersampling outperforming oversampling techniques. Evaluation metrics such as Overall Accuracy, F1-score, Precision, Recall and G-Mean score were employed, with the G-Mean score preferred for assessing individual class accuracy in imbalanced datasets. Findings indicate that algorithm level balancing with class weighted loss shows the best results, improving overall accuracy to 67.50% with a G-Mean score of 59.46%. This research advances crop mapping methodologies by providing insights into effective strategies for addressing class imbalance in deep learning-based classification, which is crucial for accurate classification of crops.
Landslides initiate when a geophysical mass, including rocks, mud, or debris, gets dislodged due to gravitational force, often triggered by various factors like rainfall, earthquakes, or anthropogenic activities. These events are particularly dangerous in mountainous areas, where they can cause significant harm to both human life and infrastructure. Detection of landslides is essential not only for diminishing damage but also for developing effective disaster management strategies. In this study, landslide susceptibility mapping was carried out in parts of the Kailash Mansarovar Pilgrimage road using multi-criteria decision-making methods like the Analytical Hierarchy Process (AHP) and the AHP-Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS). A landslide inventory was initially created by visual analysis of images sourced from Google Earth Pro, which resulted in the mapping of 100 landslide occurrences in 2024 along the road. Fourteen of the most important landslide causative/conditioning parameters, including Slope, Aspect, Stream Power Index (SPI), Drainage Density (DD), Lithology, Geomorphology, Distance from Lineament, Distance from Fault, Sediment Transport Index (STI), Topographic Wetness Index (TWI), Terrain Ruggedness Index (TRI), Normalized Difference Water Index (NDWI), Normalized Difference Vegetation Index (NDVI), and Land Use/Land Cover (LULC), were utilized as inputs in this study. The landslide susceptibility maps, thus generated using both the AHP-TOPSIS and AHP methods, indicated an increasing likelihood of landslides in the study area, thereby increasing knowledge and helping in disaster risk reduction. The AHP-TOPSIS method achieved a precision of 83.10
This study develops a hybrid hydrological forecasting framework that integrates a physically based Weighted Curve Number (WCN) method with a Long Short-Term Memory (LSTM) network to improve streamflow prediction in the Tawi watershed, Western Himalaya. The research aims to enhance the physical interpretability of deep learning models while maintaining strong predictive capability in data-scarce mountainous environments. Daily rainfall, temperature, and discharge data from 2000–2020 were used to train and validate the model at Jammu and Udhampur gauging stations. The WCN-derived runoff estimate was incorporated as an additional input feature within the LSTM architecture to account for land-use and soil-controlled hydrological responses. The hybrid model demonstrated strong performance, achieving R2 values between 0.84 and 0.86 during training and testing. Future projections for 2021–2040 indicate stable seasonal discharge patterns at Jammu, whereas Udhampur shows declining mean flows and increased low-flow sensitivity based on Flow Duration Curve analysis. By combining process-based runoff estimation with deep learning, the proposed WCN–LSTM framework improves predictive robustness and provides a transferable tool for long-term water resource planning, drought risk assessment, and climate adaptation in Himalayan River basins.