ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Share
Publications Copernicus
Download
Citation
Share
Articles | Volume X-G-2025
https://doi.org/10.5194/isprs-annals-X-G-2025-277-2025
https://doi.org/10.5194/isprs-annals-X-G-2025-277-2025
10 Jul 2025
 | 10 Jul 2025

A 40-Year Journey: Eco-Hydrological Dynamics of Haiderpur Wetlands (EHDW) with Multi-Landsat, Sentinel-2 Satellite Data using Google Earth Engine (GEE) and Machine Learning

Abhinav Galodha, Faisal Hussain, Sanya Anees, Brejesh Lall, and Shaikh Ziauddin Ahammad

Keywords: Google Earth Engine (GEE), Land Use and Land Cover (LULC), Landsat Imagery, Land Surface Temperature (LST), Urban Heat Island (UHI), Haiderpur, Sustainable Urban Planning, Modified Normalized Difference Water Index (MNDWI), Normalized Difference Vegetatio

Abstract. The Haiderpur region has undergone significant LULC changes over the decades, impacting ecological balance and urban development. However, the absence of a comprehensive, longitudinal LULC dataset has hindered detailed assessments of environmental transformations and UHI effects. This study aims to generate a thorough LULC classification dataset for Haiderpur, covering the period from 1990 to 2023. Utilizing the GEE platform along with Landsat 5, 7, and 8, Sentinel-2 multi-spectral imagery, we employed a Random Forest machine learning classifier to create a high-resolution (30 m) dataset capturing key land cover categories, including waterbodies, buildup, agriculture, bare soil, swamp vegetation, and forest. In addition to classification, the study incorporated spatial and temporal analyses using indices such as the MNDWI, NDVI, and SAVI. These indices facilitated a nuanced assessment of vegetation dynamics and water features from 1990 to 2023. The study also extends predictions of LULC into the future, projecting changes for the years 2025 and 2030. Moreover, LST variations were evaluated, highlighting significant thermal changes corresponding with LULC transformations. The dataset achieved an overall accuracy of 82%, underscoring its environmental monitoring reliability. Our findings indicate an increase in built-up areas, with corresponding impacts on thermal dynamics, while total wetland and forest areas exhibited more stability. This research confirms anthropogenic influence as a primary driver of change in the region. The Haiderpur LULC dataset aligns with other LULC resources, offering a robust tool for researchers and policymakers to support sustainable urban planning, conservation efforts, and climate change adaptation in Haiderpur and similar regions. This study contributes to SDGs, particularly Goal 6 (Clean Water and Sanitation), Goal 13 (Climate Action), and Goal 15 (Life on Land), by enhancing understanding and management of wetland ecosystems.

Share