😅Hydrological Modeling Unit 13 – Remote Sensing and GIS in Hydrology

Key Concepts and Principles

• Remote sensing involves gathering information about the Earth's surface from a distance using sensors on satellites or aircraft
• GIS (Geographic Information Systems) are computer-based tools for storing, analyzing, and visualizing spatial data
• Remote sensing data is often integrated with GIS for hydrological modeling and analysis
• Electromagnetic radiation is the foundation of remote sensing, with different wavelengths interacting differently with Earth's surface features (visible, infrared, microwave)
• Spectral signatures are unique reflectance patterns of different land cover types (vegetation, water, soil) used for classification
• Spatial resolution refers to the level of detail captured in an image, with higher resolution providing more detailed information
• Temporal resolution is the frequency at which images are acquired, important for monitoring changes over time
• Passive sensors detect naturally reflected or emitted energy, while active sensors emit their own energy and measure the returned signal (radar, lidar)

Remote Sensing Basics

• Remote sensing satellites orbit the Earth at different altitudes and inclinations, affecting coverage and revisit time
  • Geostationary satellites remain fixed above a specific location, providing continuous coverage (weather monitoring)
  • Polar-orbiting satellites follow a north-south path, covering the entire Earth at regular intervals (Landsat, Sentinel)
• Sensors capture data in different portions of the electromagnetic spectrum, each with unique properties and applications
  • Visible light (0.4-0.7 μm) is used for true-color imagery and detecting surface features
  • Near-infrared (0.7-1.4 μm) is sensitive to vegetation health and water content
  • Thermal infrared (8-14 μm) measures surface temperature and is useful for evapotranspiration studies
  • Microwave (1 mm-1 m) penetrates clouds and is used for soil moisture and surface roughness estimation (Sentinel-1, SMAP)
• Preprocessing steps are necessary to correct for atmospheric effects, geometric distortions, and sensor calibration
• Image enhancement techniques (contrast stretching, band combinations) improve visual interpretation and feature extraction

GIS Fundamentals

• GIS data is represented in either vector (points, lines, polygons) or raster (grid of cells) format
• Vector data is suitable for discrete features (rivers, watersheds), while raster data is used for continuous variables (elevation, precipitation)
• Coordinate reference systems (CRS) define the spatial reference for data, ensuring accurate alignment and analysis
  • Geographic CRS use latitude and longitude coordinates on a spherical Earth model (WGS84)
  • Projected CRS flatten the Earth's surface onto a 2D plane, preserving shape, area, or distance (UTM, State Plane)
• Attribute tables store non-spatial information associated with features, allowing for complex queries and analysis
• Spatial analysis tools include overlay (combining layers), buffer (proximity analysis), and interpolation (estimating values between points)
• Terrain analysis derives hydrologically relevant parameters from digital elevation models (DEMs), such as slope, aspect, and flow direction
• Watershed delineation identifies drainage basins and stream networks based on terrain and flow accumulation thresholds

Data Sources and Acquisition

• Landsat satellites have provided continuous multispectral imagery since 1972, with 30 m resolution and 16-day revisit time
  • Landsat 8 and 9 are the most recent missions, with improved sensors and data quality
• MODIS (Moderate Resolution Imaging Spectroradiometer) offers daily global coverage at 250-1000 m resolution, useful for large-scale monitoring
• Sentinel satellites, part of the European Copernicus program, provide high-resolution optical (Sentinel-2) and radar (Sentinel-1) data
• SRTM (Shuttle Radar Topography Mission) and ASTER GDEM are global digital elevation datasets at 30-90 m resolution
• LiDAR (Light Detection and Ranging) uses laser pulses to create high-resolution DEMs and 3D point clouds
• Precipitation data is available from ground-based networks (rain gauges), weather radar, and satellite-derived products (TRMM, GPM)
• In-situ measurements (streamflow, soil moisture) are used for calibration and validation of remote sensing estimates
• Data portals and cloud platforms (USGS EarthExplorer, Google Earth Engine) facilitate access to large archives of remote sensing data

Processing and Analysis Techniques

• Image classification assigns pixels to land cover classes based on their spectral signatures
  • Supervised classification uses training samples to guide the algorithm (Maximum Likelihood, Support Vector Machines)
  • Unsupervised classification groups pixels based on statistical patterns without prior knowledge (K-means, ISODATA)
• Change detection identifies differences in land cover or surface conditions over time
  • Post-classification comparison detects changes between independently classified images
  • Image differencing subtracts pixel values between two dates to highlight areas of change
• Vegetation indices (NDVI, EVI) quantify vegetation greenness and vigor using ratios of red and near-infrared reflectance
• Spectral unmixing estimates the fractional abundance of different land cover types within mixed pixels
• Radar interferometry (InSAR) measures surface deformation and elevation changes using phase differences between radar images
• Machine learning algorithms (Random Forests, Deep Learning) are increasingly used for complex classification and regression tasks
• Data fusion techniques combine information from multiple sensors or sources to improve accuracy and resolution (pan-sharpening, downscaling)

Applications in Hydrology

• Land use/land cover mapping provides essential input data for hydrological models, affecting runoff, infiltration, and evapotranspiration
• Evapotranspiration estimation using remote sensing-based models (SEBAL, METRIC) helps quantify water losses and irrigation requirements
• Soil moisture monitoring using microwave sensors (SMAP, SMOS) improves drought assessment and flood forecasting
• Snow cover mapping and snow water equivalent (SWE) estimation are critical for water supply forecasting in mountainous regions
• Flood mapping and monitoring using optical and radar imagery helps assess the extent and duration of inundation
• Water quality assessment using spectral indices (chlorophyll-a, turbidity) supports the management of lakes, reservoirs, and coastal waters
• Groundwater exploration using remote sensing indicators (lineaments, vegetation patterns) guides the identification of potential aquifers
• Glacier and permafrost monitoring using multi-temporal imagery informs water resource planning in cold regions

Challenges and Limitations

• Cloud cover and atmospheric effects can obscure or distort surface features, requiring careful preprocessing and cloud masking
• Trade-offs between spatial, temporal, and spectral resolution limit the ability to capture fine-scale processes and rapid changes
• Data volume and processing requirements can be significant, necessitating high-performance computing and storage solutions
• Ground truth data is essential for accuracy assessment and model calibration but can be costly and time-consuming to collect
• Spectral confusion between similar land cover types (e.g., bare soil and urban areas) can lead to classification errors
• Radar and lidar signals can be affected by surface roughness, vegetation structure, and moisture conditions, complicating interpretation
• Integrating remote sensing data with physical models requires careful consideration of scale, parameterization, and uncertainty propagation
• Socioeconomic and policy factors can limit the adoption and application of remote sensing technologies in water resources management

Future Trends and Developments

• Increased availability of high-resolution, multi-sensor data from new satellite missions (Landsat 9, Sentinel-6, NISAR)
• Advances in cloud computing and big data analytics (Google Earth Engine, AWS, Microsoft Azure) enable large-scale processing and analysis
• Machine learning and artificial intelligence techniques improve classification accuracy, change detection, and data fusion
• Integration of remote sensing with process-based models (SWAT, VIC) enhances understanding of hydrological fluxes and feedbacks
• Adoption of open data policies and standards (FAIR principles) promotes collaboration and reproducibility in research and applications
• Citizen science and crowdsourcing initiatives engage the public in data collection and validation, expanding the reach of remote sensing
• Development of low-cost, high-resolution sensors on drones and small satellites (CubeSats) increases the accessibility of remote sensing technology
• Incorporation of socioeconomic and demographic data with remote sensing supports integrated water resources management and decision-making