Spatial Data Format of the Week: Raster Image Formats for Spatial Analysis

Understanding "Images" as Spatial Data

In GIS, an image isn't just a picture—it's a georeferenced grid of measurements. This week covers the core raster image formats you'll encounter in spatial analysis, from the universal GeoTIFF to legacy formats still hiding in old data archives.

Core concept: Every pixel in a spatial raster has a location (coordinates) and a value (measurement). The format determines how that information is stored and accessed.

1. GeoTIFF (.tif, .tiff) — The Universal Raster Standard

What it is: Tagged Image File Format (TIFF) with embedded spatial reference information. The most common format for all raster spatial data.

Why GeoTIFF Dominates

A regular TIFF is just an image. A GeoTIFF adds:

• Coordinate Reference System (CRS): What projection? What datum?
• Geotransform: Where in the world is this image? (pixel size, rotation, origin)
• Bounding coordinates: Geographic extent
• NoData values: Which pixels mean "no data here"?

Result: Self-contained spatial dataset—no separate metadata files needed.

Multi-Band Structure

GeoTIFFs store multiple bands (layers) in one file:

True color imagery (RGB):

• Band 1: Red channel (0.63-0.69 µm)
• Band 2: Green channel (0.52-0.60 µm)
• Band 3: Blue channel (0.45-0.52 µm)

Landsat 8 (multispectral):

• Band 1: Coastal/Aerosol
• Band 2: Blue
• Band 3: Green
• Band 4: Red
• Band 5: NIR (near-infrared)
• Band 6-7: SWIR (shortwave infrared)
• Band 10-11: Thermal

Digital Elevation Model (DEM):

• Band 1: Elevation values (meters or feet)

Common Sources

Satellite imagery:

• Landsat (30m): USGS EarthExplorer, Google Earth Engine
• Sentinel-2 (10m): Copernicus Hub, Google Earth Engine
• NAIP (1m): USGS, state GIS portals, Google Earth Engine

Elevation data:

• SRTM (30m global): USGS, CGIAR-CSI, Google Earth Engine
• USGS 3DEP (10m U.S.): National Map, Google Earth Engine
• LiDAR-derived DEMs: State/local portals

Derived products:

• Land cover classifications (NLCD, ESA WorldCover)
• Climate grids (PRISM, WorldClim)
• Vegetation indices (NDVI, EVI)

Working with GeoTIFF in QGIS

Load:

1. Drag and drop .tif file into QGIS
2. Or: Layer → Add Layer → Add Raster Layer

Inspect properties:

• Right-click layer → Properties
• Information tab: CRS, extent, pixel size, band count, data type
• Symbology tab: Rendering (singleband gray, multiband color, paletted)
• Histogram tab: Data distribution

Common operations:

# Calculate NDVI (Raster Calculator)
("NIR@1" - "Red@1") / ("NIR@1" + "Red@1")

# Clip to study area
Raster → Extraction → Clip Raster by Extent

# Reproject
Raster → Projections → Warp (Reproject)

# Extract values at points
Processing → Raster analysis → Sample raster values

Compression Options

DEFLATE (lossless):

gdal_translate -co COMPRESS=DEFLATE -co PREDICTOR=2 input.tif output.tif

• Best for: Elevation data, scientific measurements
• PREDICTOR=2: Better compression for continuous data

LZW (lossless):

gdal_translate -co COMPRESS=LZW input.tif output.tif

• Best for: Classified data (land cover, categorical rasters)

JPEG (lossy):

gdal_translate -co COMPRESS=JPEG -co JPEG_QUALITY=85 input.tif output.tif

• Best for: Imagery where visual quality matters more than exact values
• Quality 85-95 balances size and fidelity

GeoTIFF Advantages

✅ Universal support — Every GIS software reads GeoTIFF

✅ Self-contained — Embedded CRS, no separate .prj file

✅ Multi-band — Store all spectral bands in one file

✅ Flexible compression — Choose lossless or lossy

✅ BigTIFF support — No 4GB limit

GeoTIFF Limitations

❌ Large file sizes — High-resolution imagery = multi-GB files

❌ Not web-optimized — Full download required (see COG next week)

❌ Single CRS — One projection per file

When to Use

Use GeoTIFF for:

• Desktop GIS analysis
• Satellite imagery processing
• Elevation models (DEMs)
• Archival storage
• Multi-band scientific data

2. JPEG with World File (.jpg + .jgw) — The Georeferenced Photo

What it is: Standard JPEG image + separate world file that provides spatial reference.

The World File System

A world file is a tiny text file with 6 lines defining the geotransform:

Example: aerial_photo.jgw

0.5                  ← pixel width in map units (0.5 meters)
0.0                  ← rotation term (usually 0)
0.0                  ← rotation term (usually 0)
-0.5                 ← pixel height (negative = north-up)
123456.789           ← X coordinate of upper-left pixel center
987654.321           ← Y coordinate of upper-left pixel center

File naming:

• image.jpg → image.jgw (common)
• image.jpg → image.jpgw (alternative)

Why JPEG + World File?

Advantages:

• Small file sizes — JPEG compression excellent for photos
• Universal viewing — Any image viewer opens JPEGs (non-GIS users)
• Legacy standard — Old aerial photos often distributed this way

Disadvantages:

• Lossy compression — Not suitable for analysis requiring exact values
• No CRS embedded — Requires separate .prj file for projection
• No band metadata — Typically 3-band RGB only

Common Sources

• Historical aerial photography (pre-digital era scans)
• Orthophotos (georectified aerial photos)
• Field photos with GPS cameras
• Exported web map images

Using JPEG+World in QGIS

Load:

1. Ensure three files present:
   • photo.jpg (image)
   • photo.jgw (world file)
   • photo.prj (projection, optional but recommended)
2. Drag photo.jpg into QGIS
3. QGIS automatically reads .jgw and .prj

If CRS missing:

• Right-click layer → Set CRS
• Choose appropriate coordinate system

Export with world file:

1. Project → Import/Export → Export Map to Image
2. Check "Save world file"
3. QGIS creates .jpgw automatically

When to Use

Use JPEG+World for:

• Sharing georeferenced photos with non-GIS users
• Reducing file sizes for visual display (not analysis)
• Historical aerial photos (legacy format)
• Quick web map exports

Avoid for:

• Quantitative analysis (lossy compression)
• Multispectral imagery (limited to RGB)
• Scientific measurements

3. PNG with World File (.png + .pgw) — The Lossless Web Image

What it is: Portable Network Graphics image + world file. Like JPEG+World, but lossless compression.

PNG vs. JPEG

World File Format

Same as JPEG world files:

Example: screenshot.pgw

1.0                  ← 1 meter per pixel
0.0
0.0
-1.0                 ← negative = north-up
500000.0             ← X coordinate (easting)
4000000.0            ← Y coordinate (northing)

File naming:

• image.png → image.pgw (common)
• image.png → image.pngw (alternative)

Common Uses

Web map exports:

• Export styled QGIS map as PNG
• Maintain georeferencing with .pgw
• Lossless for text labels, sharp boundaries

Screenshots of GIS software:

• Capture analysis results
• Document workflows
• Share exact visual representation

Graphics with spatial reference:

• Logos on maps
• Legends with coordinates
• Transparent overlays

Using PNG+World in QGIS

Load:

1. Ensure files present: map.png, map.pgw, map.prj
2. Drag map.png into QGIS
3. QGIS reads world file automatically

Export with world file:

1. Project → Import/Export → Export Map to Image
2. Choose PNG format
3. Check "Save world file"
4. Result: output.png + output.pgw + output.prj

Transparency Support

PNG supports alpha channel (transparency):

• Useful for overlay graphics
• Logos on web maps
• Transparent NoData areas

Create transparent NoData in QGIS:

1. Layer Properties → Transparency
2. Set NoData value or transparent color
3. Export as PNG

When to Use

Use PNG+World for:

• Lossless map exports
• Graphics with text/labels
• Transparent overlays
• Screenshots with georeferencing
• Web map images

Avoid for:

• Large imagery datasets (use GeoTIFF)
• Multispectral analysis (limited bands)
• High-color-depth scientific data

4. ESRI ASCII Grid (.asc, .txt) — The Human-Readable Raster

What it is: Plain text raster format where every pixel value is written as text. Simple, readable, inefficient.

File Structure

Example: Small 5×5 elevation grid

ncols         5
nrows         5
xllcorner     123456.0
yllcorner     987654.0
cellsize      10.0
NODATA_value  -9999
45.2 46.1 44.8 47.3 48.9
43.7 45.3 46.0 46.8 47.2
42.1 43.9 44.5 45.1 46.3
40.8 42.2 43.0 43.8 45.0
39.5 40.7 41.4 42.1 43.5

Header metadata:

• ncols / nrows: Grid dimensions
• xllcorner / yllcorner: Lower-left corner coordinates
• cellsize: Pixel size (same in X and Y)
• NODATA_value: Code for missing data

Why ASCII Grid Exists

Advantages:

• Human-readable — Open in text editor, inspect values
• Universal — Any software can parse text
• Simple — No binary format complexity
• Debugging — Easy to verify data correctness

Disadvantages:

• Massive file sizes — 10-100× larger than binary formats
• Slow I/O — Text parsing much slower than binary reads
• Limited metadata — No CRS embedded (requires .prj file)
• No compression — Every digit written explicitly

Common Sources

Modeling outputs:

• Hydrological models (flow direction, accumulation)
• Terrain analysis (slope, aspect)
• Species distribution models
• Academic research software (legacy exports)

Data interchange:

• Sharing data between incompatible software
• Educational datasets (easy to inspect)
• Manual data creation (small grids)

Using ASCII Grid in QGIS

Load:

1. Drag .asc file into QGIS
2. Or: Layer → Add Layer → Add Raster Layer
3. QGIS converts to internal format for display

Assign CRS (if missing):

1. Right-click layer → Set Layer CRS
2. Choose projection matching coordinate values

Export to ASCII:

1. Right-click layer → Export → Save As
2. Format: Arc/Info ASCII Grid
3. QGIS writes header + values

When to Use

Use ASCII Grid for:

• Quick inspection of raster values
• Data interchange between incompatible tools
• Educational examples (show actual values)
• Manual small-grid creation
• Debugging spatial analysis

Avoid for:

• Large datasets (file size explosion)
• Production workflows (too slow)
• Web delivery (inefficient)
• Archival storage (GeoTIFF better)

5. ESRI GRID — The Legacy Directory Format

What it is: Binary raster format from ESRI Arc/INFO (1980s-1990s). Directory-based, not a single file.

Structure

Directory with multiple files:

elevation/
├── dblbnd.adf       ← Bounding coordinates
├── hdr.adf          ← Header (dimensions, cell size)
├── sta.adf          ← Statistics
├── vat.adf          ← Value Attribute Table (for integer grids)
├── w001001.adf      ← Data tiles
├── w001001x.adf     ← Index files
└── info/            ← Subdirectory with metadata
    ├── arc.dir
    └── arc0000.dat

Why ESRI GRID Exists

Historical context:

• Pre-GeoTIFF era (1980s-1990s)
• ESRI's native raster format for Arc/INFO
• Designed for computational efficiency (tiled storage)
• Legacy format still encountered in old datasets

Limitations

Problems:

• Complex structure — Dozens of files in directory
• Platform-specific — Naming restrictions (8.3 DOS filenames)
• No compression — Wastes disk space
• Proprietary — ESRI format, limited non-ESRI support
• Deprecated — ESRI recommends GeoTIFF now

When You'll Encounter It

• Old projects (pre-2000 vintage)
• Legacy archives (university/government data)
• ArcGIS outputs (older workflows still use it)
• Historical datasets (elevation, land cover from 1990s)

Working with ESRI GRID in QGIS

Load:

1. Add Raster Layer → Navigate to GRID directory
2. Select directory (not individual files)
3. QGIS uses GDAL driver to read

Convert to GeoTIFF:

1. Right-click layer → Export → Save As
2. Format: GeoTIFF
3. Choose compression (DEFLATE recommended)

Command-line conversion:

gdal_translate -of GTiff -co COMPRESS=DEFLATE elevation/ elevation.tif

Recommendation

Always convert to GeoTIFF when encountering ESRI GRID:

• Single file (easier management)
• Compression (smaller size)
• Better cross-platform support
• Modern standard

Format Comparison Table

The Bottom Line

GeoTIFF is the standard for raster spatial analysis:

• Self-contained (embedded CRS, no separate files)
• Flexible compression
• Multi-band support
• Universal compatibility

World files (JPEG, PNG) are useful for:

• Sharing visual maps (non-GIS users)
• Web graphics
• Historical aerial photos

ASCII Grid is for:

• Quick inspection
• Debugging
• Data interchange (when nothing else works)

ESRI GRID is legacy—convert to GeoTIFF immediately.

See Also

• Week 6: Cloud Optimized GeoTIFF (COG) & JPEG2000 — Modern web-optimized rasters
• Week 8: NetCDF — Multi-dimensional scientific arrays
• Week 9: HDF5 — Hierarchical scientific data
• GDAL Raster Formats — Comprehensive format documentation