Search

This study investigates the potential of land use and land cover (LULC) maps in balancing preservation, development pressures, and competing uses. This study employed the Artificial Neural Networks—Cellular Automaton model to forecast changes in LULC in the Adyar basin watershed in southern India between 2000, 2010, and 2020. The model incorporates topographical variables such as DEM, slope, aspect, distance from the road, and distance from built-up land to determine the influence on LULC changes. Results show significant LULC changes to elevation and distance from the road. The model employed has demonstrated high accuracy, with a Kappa index of 0.80 and an 81% accuracy rate compared with the observed LULC maps for 2020. The study has also forecasted LULC changes for 2030, 2040, and 2050, projecting an increase in urban areas by 6.84% and a decrease of approximately 16.4% and 0.5% in agriculture and waterbody, respectively. The model has also revealed a decline of 19.1% in agriculture and 1.2% in water bodies over the years 2000 to 2050 due to human activities such as deforestation and conversion of croplands to plantations. Conclusively, the results and the interpretations support the ANN-CA model as a reliable tool for the simulation and prediction of outcomes with variables in LULC changes. [ABSTRACT FROM AUTHOR]

Coastal dune distinction is an essential for monitoring, conservation and sustainable management of fragile coastal ecosystem. Most of the coastal scientists classified the coastal sand dunes by conventional methods as frontal dunes (foredune) and back dunes (stabilized dune). Due to the significance in the developing stage of the dunes and the coastal protection, foredunes are very significant in this context. Studies demarcating the boundary of coastal dune field based on its spectral signatures of vegetation species have not been attempted by any authors earlier. In this view, the goal of this study was to determine whether it is possible to demarcate the foredune from the dune field using vegetation species with the aid of Airborne Visible InfraRed Imaging Spectrometer—Next Generation (AVIRIS-NG) hyperspectral data. To achieve this, the study scrutinized the major vegetation species and the end members were identified using Pixel Purity Index algorithm (PPI), N-Dimensional visualizer, Insitu spectra and Linear Spectral Unmixing algorithm. Identified training pixels falling in the dune field were grouped into six training classes to perform classification techniques (Spectral Angle Mapper (SAM), Spectral Informative Divergence (SID) and Support Vector Machine (SVM)). This study also discussed about various methods of selecting training pixels for demarcation of foredunes. Accuracy assessment reveals that SVM shows an overall accuracy of 91.97% compared to SID of 68.25% and SAM of 58.52% with kappa coefficient of 0.89, 0.53 and 0.49 respectively. The study found the indicator species such as Ipomoea Pes-caprae, Spinifex Littoreus and Launaea Sarmentosa blankets the foredunes in this study area, which could be used to demarcate the boundary of the foredunes. The methodology employed in this study can be applied to distinguish the foredune from the dune field in any coastal dune ecosystem, provided that the researcher has access to a database of the flora present in the area. The proposed methodology and this research might bring new insight into the classification and demarcation of foredune using the vegetation species present over the dune field. The research allows more efficient and cost-effective identification of foredunes from vast coastal environments, providing valuable information for the conservation and sustainable management of ecologically sensitive ecosystems. [ABSTRACT FROM AUTHOR]

Coastal sand dunes are the dynamic depositional landform which are formed near the coast, due to aeolian activity. Coastal sand dune acts as a natural barrier against coastal erosion hence its periodic monitoring is essential. Modification of this landform due to anthropogenic activities was observed in the coastal stretch towards the south of Tamil Nadu, even though these dune ecosystems were meant to be conserved and preserved in the coastal zone regulation notification. Present study, focuses on finding key techniques required to demarcate the morphological parameters such as ridge line, crest portion, toe, dune orientation, and its pattern from the open source ALOS World 3D (AW3D) DEM for periodic monitoring of coastal dune ecosystems. Topographic models such as hill shading, slope, aspect, curvature, and hybrid techniques were performed. The outcomes obtained from topographic models were validated with spatial profiling and field survey. From the study it is observed that, hybrid techniques particularly slope overlaid on hill shade, slope, gaussian stretched hill shade image, and spatial profiling are the key techniques to bring out the morphological variations and the parabolic dune patterns, which were hitherto not reported in earlier literatures. Gaussian stretched hill shade image enhanced the micro landforms which were difficult to extract from normal hill shading. The slope image of the study is found to be more informative in exhibiting the minor dunes and patterns compared to shaded relief image. The ridge portion of the dune, crest and limb portion and coastal plains were differentiated based on the slope value ranges. The hybrid image formed by overlaying slope and curvature data on hill-shaded image is more informative in providing most of the parabolic patterns formed by micro and meso scale dunes. Spatial profile could be used as an additional tool to validate the information's derived from topographic models, wherever there is limitation in field accessibility. [ABSTRACT FROM AUTHOR]

Image transformation techniques such as Principal Component Analysis, Minimum Noise Fraction and Independent Component Analysis were used for extracting information from the satellite image in various fields, but its application in landform mapping seems to be very less. The objective of the study is to demonstrate the viability of less explored image transformation techniques for delineating coastal dune complex and associated features using European satellite Sentinel 2 multispectral data. Based on the difference in lithology and vegetation cover the study sites has been chosen to extract coastal dune complex features such as dune ridges, dune with or without vegetation, swale with or without vegetation and water. Overall results of ICA shows enhanced visualization followed by PCA in bringing up the fine variations, identifying micro landforms within the dune complex from the optimum band combinations. Thus, sentinel data coupled with advanced image processing technique like ICA could identify maximum information about the boundaries and micro landform within the coastal sand dune complex in a cost-effective manner. However, in areas with dense vegetation cover none of these techniques has given good results in identifying the boundary or in identifying the micro landforms. [ABSTRACT FROM AUTHOR]