Geospatial data is collected through a surprisingly wide range of methods, from satellites orbiting hundreds of miles above Earth to sonar pulses bouncing off the ocean floor. The approach depends on what’s being measured: land elevation, building locations, water depth, road networks, or something else entirely. Most collection methods fall into a few major categories: satellite and aerial remote sensing, GPS positioning, laser scanning, sonar mapping, ground-based surveying, and crowdsourced data from mobile devices.
Satellite Remote Sensing
Satellites collect geospatial data using two fundamentally different types of sensors: passive and active. Passive sensors measure natural energy, typically sunlight reflected off Earth’s surface or heat radiated from it. The specific wavelengths a passive sensor detects reveal information about surface composition, temperature, roughness, and vegetation health. These sensors work day and night in all weather conditions, making them ideal for continuous global monitoring. Programs like Landsat and Sentinel use passive optical sensors to produce the satellite imagery most people are familiar with.
Active sensors work differently. They transmit their own signal (usually radar or microwave pulses) toward the ground and measure what bounces back. A precipitation radar, for example, reads the echo from rainfall to calculate how fast rain is falling across a region. Cloud-profiling radars build three-dimensional maps of cloud structure. Because active sensors generate their own energy, they can penetrate cloud cover and work in complete darkness, which gives them a significant advantage in tropical or polar regions where clear skies are rare.
GPS and Satellite Positioning
The Global Positioning System (and its international counterparts like Europe’s Galileo and Russia’s GLONASS) collects location data through a method called trilateration: measuring the distance between a receiver and multiple satellites whose positions are already known. In theory, three distance measurements are enough to pinpoint a location in three dimensions. In practice, GPS requires signals from at least four satellites. The fourth signal is needed to correct for timing errors in the receiver’s internal clock, which would otherwise throw off the distance calculations by significant amounts.
Each satellite broadcasts a navigation message that lets the receiver compute the satellite’s exact position in space and its clock offset. The receiver then solves a system of equations with four unknowns: its own three coordinates (latitude, longitude, and elevation) plus its clock error. More satellites in view means more equations and a more accurate fix. Modern receivers in open sky conditions routinely achieve accuracy within a few meters, while professional-grade equipment using correction techniques can reach centimeter-level precision.
LiDAR and Laser Scanning
LiDAR (Light Detection and Ranging) collects geospatial data by firing rapid laser pulses toward the ground and measuring how long each pulse takes to return. That round-trip travel time is converted into a distance, which is then translated into an elevation value. But knowing the distance alone isn’t enough. A LiDAR system also relies on an onboard GPS to track the exact position of the aircraft or drone, and an Inertial Measurement Unit that records the platform’s orientation (its roll, pitch, and yaw) at the moment each pulse is fired. Combining all three measurements, the system assigns precise 3D coordinates to every point the laser hits.
The result is a “point cloud,” a dense collection of millions of individual elevation measurements that together form a detailed 3D model of the landscape. LiDAR systems record these returns in one of two ways. Discrete return systems identify the strongest peaks in the reflected light energy and record a separate point for each peak. A single laser pulse passing through a forest canopy might generate anywhere from 1 to 11 or more discrete returns as it bounces off leaves, branches, and finally the ground. Full waveform systems take a different approach, recording the entire distribution of returned light energy as a continuous curve. This captures more subtle details about the vertical structure of whatever the laser passed through.
LiDAR data quality is held to strict accuracy standards. The current national standards define vertical accuracy in terms of root mean square error. The highest-quality data (QL0) must achieve vertical accuracy of 5 centimeters or better in open terrain, while standard mapping-grade data (QL2) allows up to 10 centimeters. These thresholds ensure the resulting elevation models are reliable for floodplain mapping, infrastructure planning, and other applications where precision matters.
Photogrammetry and Aerial Imagery
Photogrammetry turns overlapping 2D photographs into 3D geospatial data. The core idea is simple: if you photograph the same area from multiple angles, software can calculate the depth and position of features by comparing how they shift between images. The modern version of this technique, called Structure from Motion, automates most of the process. SfM algorithms analyze overlapping photos and automatically estimate the camera’s internal geometry, position, and orientation for each shot based on the image data alone, without needing detailed flight path records or camera calibration files.
The output is a 3D point cloud similar to what LiDAR produces, along with orthorectified mosaics (stitched aerial photos corrected for distortion so they can be used as maps). A small number of Ground Control Points, physical markers on the ground with known coordinates, are used to anchor the model to real-world locations. This method works with everything from professional aerial survey cameras to consumer drones, which has made high-resolution 3D mapping far more accessible in recent years.
Sonar and Seafloor Mapping
Underwater geospatial data relies on sound instead of light. Multibeam sonar systems, mounted on the hull of a survey vessel, send out multiple simultaneous sound pulses in a fan-shaped pattern beneath the ship. By measuring the time each pulse takes to travel to the seafloor and return, the system calculates the water depth at each point. Scientists onboard also measure the speed of sound in the local water column (which varies with temperature, salinity, and pressure) so they can convert travel times into accurate depth values.
Unlike older single-beam sonar that mapped one narrow strip directly below the vessel, multibeam systems cover a wide swath of seafloor with each pass. They also collect backscatter data: information about how strongly the seafloor reflects the sound. Hard surfaces like rock produce strong backscatter, while soft sediment absorbs more energy. Together, the depth and backscatter data create detailed maps of both the shape and composition of the ocean floor.
Ground-Based Surveying
Traditional land surveying remains one of the most accurate ways to collect geospatial data for specific points. Total stations, the workhorse instruments of the surveying profession, combine an electronic distance meter with a precision angle-measuring device. A surveyor aims the instrument at a reflective target on a rod, and the total station measures both the distance and the horizontal and vertical angles to that point, calculating exact coordinates. This method is especially useful in areas where GPS signals are blocked or unreliable, such as urban canyons, dense forests, or inside buildings.
Real-Time Kinematic GPS has become a common complement to total stations. RTK uses a base station at a known location to broadcast corrections to a nearby rover unit, pushing GPS accuracy from meters down to one or two centimeters. Surveyors typically use RTK for open-area work where speed matters and switch to a total station when they need line-of-sight measurements in obstructed environments.
Crowdsourced and Mobile Data
Smartphones and connected devices have become a massive source of geospatial data. GPS traces from commuters’ mobile devices generate real-time traffic updates for navigation apps. Geo-tagged audio samples recorded by pedestrians can be aggregated to build citywide noise pollution maps at different times of day. Location-aware apps collect everything from pothole reports to disease outbreak locations, turning millions of everyday users into a distributed sensor network.
This category, sometimes called Volunteered Geographic Information, powers platforms like OpenStreetMap and feeds into public health surveillance, emergency response, and urban planning. The tradeoff is accuracy. A smartphone GPS fix is typically accurate to about 3 to 5 meters under good conditions, which is fine for mapping a road or tagging a photo but not for boundary surveys or engineering work.
How the Data Is Stored
Once collected, geospatial data is stored in one of two fundamental formats: vector or raster. Vector data represents features as points, lines, or polygons, and it works best for things with defined boundaries or exact locations. County borders, road centerlines, fire hydrant locations: these are all vector data. The most common vector file format is the shapefile, which is actually a bundle of at least three files (.shp, .shx, and .dbf) that must stay together to function.
Raster data represents continuous surfaces as a grid of cells, where each cell holds a single value. Elevation, temperature, precipitation, and soil chemistry are natural fits for raster format because they vary smoothly across space rather than having hard edges. Satellite imagery is also stored as raster data. Common raster formats include GeoTIFF (a standard image file with geographic coordinates embedded in the header), Digital Elevation Models used by the USGS for terrain data, and specialized remote sensing formats like BIL, BIP, and BSQ. The choice between vector and raster isn’t about quality. It’s about what kind of information you’re representing and what you plan to do with it.