Raster Data and Remote Sensing Basics

What Is Raster Data?

A raster is a grid of equally spaced cells, sometimes called pixels, where every cell holds one or more numeric values. Think of a digital photograph: rows and columns of pixels, each storing a color value. Spatial rasters work the same way, but every cell also covers a specific patch of Earth's surface. Two rasters can describe the same area but look very different if one has 30-meter cells and another has 1-meter cells — that difference is called spatial resolution.

Rasters store values in bands. An elevation raster (Digital Elevation Model, or DEM) has a single band: each cell stores the ground height in meters or feet. A Landsat 8 satellite scene has multiple bands — separate grids for blue light, green light, red light, near-infrared, and several others. By combining bands you can calculate indices such as NDVI (Normalized Difference Vegetation Index) that reveal vegetation health across a landscape.

Every raster has an extent: the bounding rectangle (minimum and maximum X and Y coordinates) that defines the geographic area the dataset covers. Cells outside collected data carry a special placeholder called a NoData value — commonly −9999 or −32768 — that tells the software to ignore those locations during analysis. Always check the NoData setting before you begin analysis; treating NoData as real zeros will corrupt your statistics.

The USGS National Elevation Dataset (NED) DEM

The USGS National Elevation Dataset provides bare-earth elevation data across the United States at several resolutions. The 1/3 arc-second product (roughly 10-meter horizontal spacing at mid-latitudes) is freely downloadable from USGS EarthExplorer. When you load a NED tile in QGIS, the single raster band stores elevation in meters relative to mean sea level. Ridges appear as high values; valleys and floodplains appear as low values.

Landsat 8 Multispectral Data

Landsat 8, operated by USGS and NASA, collects imagery in multiple spectral bands at 30-meter resolution for most bands. Band 4 captures red light; Band 5 captures near-infrared. A color composite that maps any three bands to the red, green, and blue display channels lets you visualize surface features. A false-color composite using near-infrared, red, and green makes healthy vegetation appear bright red and is a standard tool for land-cover interpretation.

Reclassification

Reclassification assigns new, simplified values to ranges of original cell values. For an elevation raster you might group all cells below 200 m into class 1 (low), cells from 200–500 m into class 2 (medium), and cells above 500 m into class 3 (high). This turns a continuous dataset with thousands of unique elevation values into a categorical map with just three classes — much easier to display and communicate.

In QGIS, the Reclassify by table tool (found under Raster → Raster Analysis, or through the Processing Toolbox) lets you enter minimum, maximum, and new-value rows for each class. Set the output data type to integer when working with simple categorical classes to keep file sizes small.

Zonal Statistics

Zonal statistics summarize raster values within polygon boundaries. The polygons define zones (for example, census tracts or watershed boundaries). For each zone the tool calculates statistics such as mean, minimum, maximum, sum, and count of raster cells that fall inside that polygon. The result is written back to the polygon attribute table as new fields.

You might overlay watershed boundaries on an NDVI raster to find out which watershed has the highest average vegetation density, or overlay census tracts on a DEM to identify which neighborhood sits at the highest average elevation. Those mean values become the numbers you cite in your written interpretation.

In QGIS, the Zonal Statistics tool is in the Processing Toolbox under Raster Analysis. You select the raster layer, the zone polygon layer, choose a field-name prefix, and check the statistics you want. The tool adds columns such as `_mean`, `_min`, and `_max` to the polygon attribute table.

File Formats

The standard output format for raster analysis in this course is GeoTIFF (.tif). A GeoTIFF embeds coordinate reference system (CRS) information directly in the file header, so any GIS software can read its location without a separate world file. When you export your classified raster, confirm the CRS matches your project CRS and that the correct NoData value is written into the file.

Data Safety and System Safety

Raster files — especially full Landsat scenes — can be hundreds of megabytes to several gigabytes. Before processing, check that your workstation has enough free disk space (aim for at least three times the raw file size as working space) and save your QGIS project file frequently. Never store project data only on a USB drive that could be unplugged mid-process; use a local drive or a backed-up network location.

Always verify the coordinate reference system before overlaying raster and vector layers. Mismatched CRS settings are the most common source of incorrect zonal statistics — polygons that look aligned on screen may not actually be snapped to the correct cells if projections differ. Use QGIS's "Reproject" tools to put all layers in the same CRS before analysis.

Workflow Summary

1. Download and load the NED DEM or Landsat scene into QGIS; inspect metadata (CRS, resolution, NoData value).

2. Stretch the display symbology to confirm data loaded correctly.

3. Run Reclassify by table to create a three-class elevation raster; export as GeoTIFF.

4. Load your polygon zone layer; confirm it shares the same CRS as the raster.

5. Run Zonal Statistics; open the attribute table and record mean values per zone.

6. Write your interpretive paragraph citing at least two specific numeric values from your output.