Urban Heat Island Mapping with ArcGIS and Remote Sensing

Remote sensing is a science and technology of obtaining information about the Earth's surface without direct physical contact. It involves collecting information about areas or objects from a distance using sensors mounted on vehicles, aircraft, or satellites. The fundamental principles include: (1) Energy source (typically the sun) emits radiation, (2) Radiation passes through the atmosphere, (3) Radiation interacts with the Earth's surface (some absorbed, some reflected), (4) Sensors detect the reflected radiation, (5) Data is transmitted and recorded, (6) Image processing occurs, and (7) Final interpretation is performed. Remote sensing uses electromagnetic radiation (EMR) across the spectrum: Gamma rays (shortest wavelength, highest frequency, medical therapy), X-rays (medical imaging), Ultraviolet (fluorescent sensors), Visible light (human vision and photography), Near infrared, Short wave infrared, Thermal infrared (agriculture for soil moisture, mineral mapping, temperature), Microwaves (radar and weather systems), and Radio waves (communication, TV, FM, navigation). Remote sensing is categorized into Passive (uses natural energy sources like the sun) and Active (sensors generate their own energy, such as radar and lidar). Remote sensing images have four resolution types: (1) Spatial resolution (smallest ground area sensor can detect, determining image clarity); (2) Spectral resolution (ability to distinguish different wavelengths, allowing material differentiation); (3) Temporal resolution (time interval between observations: daily, every 5 days, every 16 days, monthly); (4) Radiometric resolution (detecting small differences in reflected energy/brightness). Satellite types include: Optical (visible light, natural color images, requires daylight/clear weather), Multispectral (2-15 bands), Hyperspectral (more bands for detailed material differentiation), Thermal infrared (detects heat, operates day/night), and Radar (microwave, active sensing, operates through clouds/rain). Remote sensing differs from photogrammetry (3D models from aerial photographs) and aerial survey (aircraft/drones for direct image capture). Applications span agriculture monitoring, forest management, water resource management, disaster management, and urban planning.

Remote sensing involves capturing electromagnetic radiation (EMR) from Earth's surface using sensors that detect various wavelengths including visible light (blue, green, red), near infrared, and thermal infrared bands; active sensors emit microwave signals and capture reflected energy, while passive sensors detect naturally emitted or reflected radiation, with images composed of pixels representing intensity values at specific locations on Earth's surface.

Remote sensing measures objects without physical contact, distinguishing active systems (emitting signals like CloudSat) from passive systems (receiving natural radiation like MODIS). The electromagnetic spectrum determines satellite capabilities: solar radiation peaks at 0.5 micron (visible), Earth's thermal emission at 10 micron. Atmospheric windows (like 0.58-0.68 micron) enable surface viewing by minimizing absorption by water vapor, CO2, and O3. Planck's Law describes spectral distribution, while Wien's Law explains why 4-micron bands detect forest fires (peak emission at 3.9-4.0 microns for 800-1000K fires). Different Earth features have distinct spectral fingerprints: healthy vegetation shows high near-infrared reflectance, water reflects less than 10%, and clouds appear spectrally flat. Sensor performance depends on spatial resolution (15m to 1km), spectral resolution (bandwidth and number of bands), temporal resolution (15 minutes to daily), and radiometric resolution (8-bit to 12-bit). False color composites using infrared wavelengths enhance feature separation—dust appears pink, volcanic ash differs from ice clouds in thermal infrared bands (10-12 microns). Image classification separates pixels into thematic classes using supervised (training samples, maximum likelihood) or unsupervised (clustering) methods.

Remote sensing relies on the electromagnetic spectrum—the complete range of energy moving at light speed, originating from the Sun. Visible light represents only 400-700 nanometers; remote sensing primarily uses infrared wavelengths beyond human vision. When solar energy reaches Earth, it interacts with matter through transmission, absorption, reflection, scattering, or emission. Remote sensing sensors function like specialized cameras, measuring radiation within specific wavelength ranges called bands. Each band produces a grayscale image encoded as digital numbers (0-255). Landsat satellites exemplify multispectral systems with 8-12 bands covering visible and infrared wavelengths. The panchromatic band captures broader wavelengths for higher resolution (15m) but lacks color discrimination. Near-infrared bands are critical for vegetation analysis since plants strongly reflect this wavelength.

Electromagnetic radiation is the fundamental tool for thermal remote sensing. Light exhibits wave-particle duality: wave properties explain reflection, refraction, interference, and polarization, while particle properties explain the photoelectric effect. Electromagnetic waves are self-propagating with electric and magnetic components oscillating at 90 degrees. Key properties include speed (3×10^8 m/s in vacuum), wavelength, frequency, phase, and amplitude. All matter above 0 Kelvin emits radiation following the Stefan-Boltzmann law (M = σT^4), where higher temperatures produce higher frequency, shorter wavelength radiation. The electromagnetic spectrum spans from radio waves to visible light (400-700 nm) to ultraviolet (10-400 nm) to infrared. Thermal infrared (3,000-35,000 nm) is critical for remote sensing, with atmospheric windows (8,000-14,000 nm) allowing satellite observation of Earth's emitted radiation for temperature estimation.

Land Surface Temperature (LST) is a radiative temperature measure of how hot or cold the Earth's surface would feel to the touch, differing from air temperature measured at 2 meters height. LST can be measured using airborne, spaceborne, or ground-based instruments. It is an essential climate variable that evaluates land-atmosphere exchange processes and constrains surface energy budgets. The split window approach uses two adjacent channels (11 and 12 microns) for atmospheric correction. Infrared retrieval offers stronger signals, higher emissivity, and better spatial resolution (1 km to 100m), while microwave provides cloud penetration but higher uncertainty. Surface emissivity ranges from 0.94 for sandy soils to over 0.99 for water, snow, ice, and vegetation. Four validation categories exist: in-situ validation, radiance-based validation, intercomparison with other products, and time series analysis. Effective validation requires homogeneous land cover, accurate geolocation, and accounting for viewing angle effects. Viewing angle differences between sensors can cause differences exceeding 5 Kelvin at swath edges.

Satellite data measures Land Surface Temperature (LST), which is the temperature radiated by the land surface, generally much hotter than air temperature. If humans were to lie on black asphalt in extreme heat areas, they would burn quickly, demonstrating the significant difference between what we feel as air temperature versus actual surface temperatures.

Land Surface Temperature (LST) and Air Temperature are two distinct temperature measurements. LST is measured at the ground surface, while Air Temperature is measured at a standard height of 1.2 meters above the ground. Although there is a strong relationship between them, they have different physical meanings and respond differently to atmospheric conditions.

There are two main types of temperature measurements: Land Surface Temperature (LST) from satellite sensors (measuring radiometric temperature of the surface) and Air Temperature (T2m) from reanalysis data or stations (measuring temperature at 2 meters height). These are fundamentally different measurements and should not be directly compared without understanding their physical differences.

Land surface temperature (LST) differs fundamentally from air temperature. LST represents the temperature of the actual ground surface, including rooftops, pavement, and other built surfaces, rather than the air temperature experienced by humans. This distinction is critical for understanding urban heat islands, as LST captures how surfaces absorb and retain heat, while air temperature reflects atmospheric conditions.

This lesson teaches how to download free Landsat satellite imagery from USGS EarthExplorer for remote sensing applications. Landsat 7 (1999) and Landsat 8 (2013) provide multi-spectral images with 30-meter resolution (15-meter for panchromatic band) and revisit every 16 days. The process involves: (1) registering on EarthExplorer, (2) defining the study area using coordinates or shapefiles, (3) selecting the desired Landsat mission and spectral bands, (4) specifying the temporal range, (5) previewing available images to verify coverage, and (6) downloading the georeferenced data (11 bands for Landsat 8, each ~120MB except the panchromatic band at ~500MB). The downloaded files are compressed and require decompression before use in ArcGIS for analysis.

The Landsat program provides the longest continuous record of Earth's land surface since 1972, managed jointly by NASA and USGS. NASA handles satellite design and launch while USGS manages ground operations and archives. Landsat 8 features the Operational Land Imager (OLI) with 9 spectral bands and Thermal Infrared Sensor (TIRS) with 2 thermal bands, including innovative Band 9 for cirrus cloud detection. Remote sensing resolution encompasses spectral (wavelength ranges), spatial (30m pixel size), temporal (16-day revisit cycle), and radiometric parameters. Multiple access platforms exist: USGS Earth Explorer for comprehensive searching, GloVIS for browse viewing, Landsat Look Viewer for interactive exploration, and NASA Reverb. Data products range from standard Level 1 multispectral data to specialized Climate Data Records providing atmospherically corrected surface reflectance for long-term environmental monitoring.

This comprehensive segment covers the entire workflow for downloading free Landsat satellite imagery from USGS Explorer. Users learn to access earthexplorer.usgs.gov, define geographic search areas using interactive maps, select acquisition dates and cloud cover thresholds, choose appropriate Landsat collections (Collection 1/2) and processing levels (Level 1/2), and understand sensor systems (Landsat 8-9 OLI, Landsat 7 ETM+, Landsat 5 TM). The video explains account requirements, metadata interpretation, spectral band selection, and download options. Users learn to interpret image metadata including collection, level, row/column coordinates, and acquisition dates. The segment concludes with initiating downloads, managing download processes, and extracting compressed files for subsequent processing.

The Landsat program, a joint NASA and USGS initiative since 1972, provides multispectral satellite imagery. Landsat 8 and 9 have 11 spectral bands, with bands 2 (blue), 3 (green), and 4 (red) being most commonly used for visual analysis. The panchromatic band (band 8) collects light from a wider wavelength range, enabling higher resolution. For agricultural analysis, near infrared and shortwave infrared bands are valuable. All bands have 30-meter resolution except the panchromatic band. The Earth Explorer website allows users to search, filter by cloud cover, and download Level 1 products (which include panchromatic bands) for GIS applications.

Landsat satellites capture Earth's surface using multispectral imaging, detecting different wavelengths of reflected sunlight. Landsat 7 has 8 bands at 30-meter resolution, while Landsat 8 has 11 bands with improved spectral coverage including band 10 for geological analysis. Different geological features can be identified through specific RGB band combinations: natural color uses bands 4,3,2; vegetation health uses 5,6,2; and mineral alteration detection requires specialized band combinations. The USGS Earth Explorer platform provides free access to Landsat imagery, requiring account creation. Key selection criteria include minimal cloud cover, adequate coverage of the study area, and temporal proximity to the research period. Level 1 data (L1T) is freely available and includes all 11 bands. After downloading, users must extract the compressed package to access individual band files for analysis.