Lab 1: The Local Baseline
Manual Landsat 8 Processing in ArcGIS Pro
Background
On the night of September 7, 2020, the Labor Day windstorm tore through the Cascades and eastern foothills of Washington State. Within 24 hours, more than 300,000 acres burned — the most destructive fire event in Washington’s recorded history. The Babb, Pearl Hill, and Cold Springs fires merged and exploded across the landscape at speeds that outpaced every evacuation route.
From the ground, the devastation was chaotic. From space, it was remarkably legible.
That contrast — the difference between what a human witness sees and what a sensor records — is the core insight of this course. This week you will work with a Landsat image of the burn scar acquired in the weeks after the fires. You will download the data yourself, manually assemble band combinations in ArcGIS Pro, and read pixel values directly from the image. Every step is intentionally hands-on and procedurally explicit.
Why start here, before Google Earth Engine? Because understanding what GEE is doing on your behalf later requires having done it by hand first. By the end of this lab you will understand why a pixel glows red in one composite and is nearly invisible in another — and you will be able to trace that difference back to the physics you read about in Campbell Chapters 1 and 2.
Part 0 — ArcGIS Pro Orientation
ArcGIS Pro is a desktop GIS application made by Esri. Unlike web-based tools, it runs locally on your computer and gives you direct, hands-on control over data files. This is both its strength and its limitation — you have complete flexibility, but you are responsible for managing your own files.
The ArcGIS Pro Interface
When you first open ArcGIS Pro and create a new project, you will see four main areas:
1. The Map View (center) — This is your canvas. Spatial data you add will appear here. You can pan by clicking and dragging, zoom with the scroll wheel, and right-click for context options.
2. The Contents Pane (left) — Lists all layers currently in your map, in display order. Layers at the top of the list draw on top of layers below. You control visibility (eye icon), layer order (drag), and symbology (double-click the color swatch) from here.
3. The Catalog Pane (right) — Your file browser. Connects to folders, databases, and online sources. You can drag data directly from here into the Map View.
4. The Ribbon (top) — Tabbed menu of tools organized by task. The Map tab has navigation tools; the Analysis tab has geoprocessing; the Insert tab lets you add map layouts.
Creating a New Project
- Open ArcGIS Pro → click New → select Map
- Name your project
Lab01_YourLastNameand save it to your designated course folder (e.g.,H:\ESCI442\Lab01\) - ArcGIS Pro creates a project folder containing a
.aprxproject file along with default geodatabase and toolbox files — do not rename or move these
Adding Raster Data
To add a raster (image) file: in the Catalog Pane, navigate to the folder containing your file → right-click the file → Add to Current Map. Or drag it directly into the Map View.
You can also use the menu: Map tab → Add Data (the yellow plus icon).
The Coordinate System
Before working with spatial data, always check your map’s coordinate system. Right-click the Map entry at the top of the Contents Pane → Properties → Coordinate System. This shows the projection currently in use. ArcGIS Pro will often auto-assign a coordinate system based on the first layer you add — this is usually correct but always worth confirming.
Part 1 — Data Acquisition
The data you will use in this lab is Landsat 8 OLI Level-2 Surface Reflectance from the USGS Landsat archive. “Level-2” means the raw satellite signal has already been converted to surface reflectance — a physically meaningful quantity that represents the proportion of incoming solar radiation reflected by the ground surface. You will read more about radiance versus reflectance in Campbell Ch. 2; for now, understand that surface reflectance is the standard for scientific analysis because it allows meaningful comparisons across dates, sensors, and locations.
Step 1.1 — Log Into EarthExplorer
Navigate to earthexplorer.usgs.gov and log in. If you do not yet have an account, click Register and create one — it is free and takes about two minutes.
Step 1.2 — Define Your Search Area
You will search for imagery over the Pearl Hill Fire burn scar, located in Douglas and Okanogan Counties in north-central Washington State.
In the Search Criteria tab (left panel):
- Click the Address/Place tab within the search box
- Type:
Coulee City, Washingtonand click Show - The map will center near Coulee City, which sits close to the southern edge of the Pearl Hill burn perimeter.
- Then click the Coordinates tab and enter the following bounding box in decimal degrees (type the numbers exactly as shown — no degree symbols needed):
| Corner | Latitude | Longitude |
|---|---|---|
| Upper Left | 47.95 | -119.80 |
| Lower Right | 47.45 | -119.10 |
Step 1.3 — Set the Date Range
In the Date Range fields, enter:
- Start Date: 09/20/2020
- End Date: 11/15/2020
This window captures imagery after the fires were contained (early September) but before significant snow cover. You want to see the burn scar clearly, not smoke or clouds.
Step 1.4 — Select the Dataset
Click Data Sets >> to advance to the dataset selection panel.
Expand the following tree: Landsat → Landsat Collection 2 Level-2
Check the box for Landsat 8-9 OLI/TIRS C2 L2
Then click Results >>.
Step 1.5 — Evaluate and Download a Scene
EarthExplorer will return thumbnail images of available scenes. You are looking for a scene that:
- Covers the fire area (the burn scar appears as a large reddish-brown area east of the Cascades ridge in the browse thumbnail)
- Has low cloud cover (aim for < 15%)
- Was acquired in October or early November 2020
The scene you are looking for is on WRS-2 Path 045, Row 027 — this is the orbital grid cell that covers the Pearl Hill burn area. If results seem to show the wrong location, confirm this path/row is represented in your results.
You may find either a Landsat 8 (LC08) or Landsat 9 (LC09) scene — both are acceptable for this lab. The band structure and file naming convention are identical between the two sensors.
Select the Product Bundle option — this downloads all bands plus metadata as a single compressed .tar file (approximately 1–2 GB).
Step 1.6 — Extract the Files
Use 7-Zip, WinRAR, or Windows’ built-in extraction to unzip the .tar file. Place the extracted folder inside your Lab01 project directory, e.g.:
H:\ESCI442\Lab01\LC08_L2SP_045027_20201003\
📸 Screenshot 1: Take a screenshot of EarthExplorer showing your selected scene’s thumbnail overlaid on the map, with the scene metadata (path/row, date, cloud cover) visible in the results panel. Include this in your lab submission.
Part 2 — Loading Data in ArcGIS Pro
Step 2.1 — Create a Composite Raster
Rather than loading six individual band files separately, you will combine them into a single multi-band composite using the Composite Bands geoprocessing tool. This creates one .TIF file containing all bands in order — easier to work with and the required input for the Spectral Profile chart later.
- In the ArcGIS Pro ribbon, go to Analysis tab → Tools (opens the Geoprocessing pane)
- In the search bar, type
Composite Bandsand click the result - In the Input Rasters field, add the bands in this order:
SR_B2, SR_B3, SR_B4, SR_B5, SR_B6, SR_B7(Blue, Green, Red, NIR, SWIR1, SWIR2) - Set the Output Raster location to your Lab01 folder and name it
L8_composite_6band.TIF - Click Run
When complete, right-click the output in the Catalog Pane → Add to Current Map.
Step 2.2 — Initial Display and Coordinate System
When the composite loads, ArcGIS Pro will likely display it as a grayscale image or with an automatic stretch. Before adjusting the display:
- Right-click Map at the top of the Contents Pane → Properties → Coordinate System
- Note the coordinate system. Landsat 8 data is delivered in UTM (Universal Transverse Mercator), projected using the WGS 84 datum. Your scene is in UTM Zone 10N.
- Click OK
📸 Screenshot 2: Take a screenshot of ArcGIS Pro with your 6-band composite loaded and the coordinate system properties dialog open, showing the UTM Zone 10N projection. Your screenshot should look similar to the example below.
Part 3 — Band Combinations and Visual Interpretation
This is the conceptual core of the lab. A satellite image is not a photograph — it is a set of independent measurements of reflected energy at different wavelengths. By choosing which wavelength measurements to display as Red, Green, and Blue on your screen, you create a composite image that emphasizes different surface properties.
Your eye never sees NIR radiation. But by mapping NIR reflectance to the red channel on your monitor, you can make the invisible visible.
Step 3.1 — Natural Color Composite (Composite Bands 3-2-1 as R-G-B)
This composite approximates what your eye would see from an airplane.
- In the Contents Pane, double-click the color swatch (thumbnail) of your composite layer to open the Symbology Pane
- Change Symbology Type to RGB
- Set:
- Red channel → Band 3 (this is Landsat B4, Red, 0.64–0.67 µm)
- Green channel → Band 2 (this is Landsat B3, Green, 0.53–0.59 µm)
- Blue channel → Band 1 (this is Landsat B2, Blue, 0.45–0.51 µm)
- Under Stretch Type, select Percent Clip. Two fields will appear for Min and Max. The right value depends on which composite you are displaying:
- For natural color (visible bands): start with Min: 0.5, Max: 0.5
- For NIR and SWIR composites: use Min: 0.2, Max: 0.2 — the NIR band has a much wider dynamic range and aggressive clipping will blow out bright surfaces to pure white
- Click anywhere outside the pane to apply
Zoom into the burn scar area (northeast of the Cascades ridge). The burn scar appears as a pale gray to whitish area in the upper right of the scene — subtle in natural color, which is intentional. Keep this view in mind as you move to the next composite.
❓ Question 1 (Interpretation): Examine the natural color composite. Identify and describe three distinct land cover types you can visually distinguish in the image (e.g., the burn scar, agricultural fields, water). For each, describe its color and suggest a reason, based on Campbell Ch. 2’s discussion of spectral response, why that surface looks the way it does in visible wavelengths.
Step 3.2 — NIR False Color Composite (Composite Bands 4-3-2 as R-G-B)
Now you will display the near-infrared band in the red channel of your monitor.
- Return to the Symbology Pane
- Set:
- Red channel → Band 4 (this is Landsat B5, NIR, 0.85–0.88 µm)
- Green channel → Band 3 (this is Landsat B4, Red, 0.64–0.67 µm)
- Blue channel → Band 2 (this is Landsat B3, Green, 0.53–0.59 µm)
- Keep Stretch Type set to Percent Clip — adjust Min/Max visually as needed (0.2–1.0 is a typical range for this data)
Healthy vegetation will now appear in bright red. The dense Cascades forests on the western side of the scene will be the most intensely red areas. The burn scar will appear as a large dusky pink to muted brownish-pink area in the upper-right portion of the scene — clearly distinguishable from the surrounding bright red vegetation. Water (the Columbia River, Banks Lake) will appear very dark navy to black.
❓ Question 2 (Interpretation): Compare the NIR false color composite to the natural color composite you made in Step 3.1. Identify at least two land cover types that are easier to distinguish in the NIR composite than in the natural color view. Explain why the NIR band makes these distinctions clearer, drawing explicitly on the concept of spectral reflectance from Campbell Ch. 2.
❓ Question 3 (Interpretation): In the NIR false color composite, find an area that appears bright red and an area that appears dark or black. Use the Identify tool (Map tab → Explore → Identify) to click on each area and record the pixel values for composite Band 3 (Landsat B4, Red), Band 4 (Landsat B5, NIR), and Band 5 (Landsat B6, SWIR1). Present these values in a small table and explain what the difference in NIR reflectance tells you about the surface condition at each location.
Step 3.3 — SWIR Composite for Burn Severity (Composite Bands 6-4-2 as R-G-B)
Short-wave infrared radiation penetrates atmospheric smoke and haze, and is highly sensitive to soil moisture and char. This composite is used operationally by fire agencies to map burn severity.
- Return to Symbology Pane
- Set:
- Red channel → Band 6 (this is Landsat B7, SWIR2, 2.11–2.29 µm)
- Green channel → Band 4 (this is Landsat B5, NIR, 0.85–0.88 µm)
- Blue channel → Band 2 (this is Landsat B3, Green, 0.53–0.59 µm)
- Keep Stretch Type set to Percent Clip — adjust Min/Max visually as needed (0.2–1.0 is a typical range for this data)
Healthy vegetation will appear bright green (strong NIR signal in the green channel). The burn scar will appear as a muted brownish area in the upper right of the scene — less dramatic than in the CIR composite, but spatially detailed. You may be able to distinguish patches of different burn intensity within the scar. Water will appear dark navy to black.
❓ Question 4 (Interpretation): Looking at the SWIR composite, do you notice any spatial patterns within the burn scar itself — areas that appear more severely burned versus less severely burned? Describe what you observe. Based on what you have read in Campbell Ch. 2 about how different surfaces produce different spectral responses, suggest one physical reason why different parts of a burn scar might show different SWIR reflectance values.
📸 Screenshot 3: Take a screenshot showing all three composites (natural color, NIR false color, SWIR) side by side, or as three sequential screenshots. Label each clearly with its band combination. Include these in your submission.
Part 4 — Spectral Profiles
So far, you have been comparing composites visually. Now you will measure the spectral signatures of specific pixels — plotting reflectance against wavelength for individual ground locations, just as Campbell Ch. 2 describes in Figure 2.19.
Step 4.1 — Download the Sample Point Shapefile
Download the instructor-provided shapefile lab01_sample_points.zip from Canvas and extract it into your Lab01 project folder. Add the shapefile to your ArcGIS Pro map (Catalog Pane → navigate to folder → drag into map).
The shapefile contains 10 sample points representing the following cover types:
| Point ID | Cover Type | Longitude | Latitude |
|---|---|---|---|
| P1 | Dense healthy forest (unburned) | -120.692117 | 47.922775 |
| P2 | Dense healthy forest (unburned) | -121.227722 | 47.19976 |
| P3 | High-severity burn | -119.37504 | 48.099905 |
| P4 | High-severity burn | -119.573736 | 47.673221 |
| P5 | Agricultural field (dry/fallow) | -119.935169 | 47.593621 |
| P6 | Agricultural field (irrigated) | -119.755401 | 47.069682 |
| P7 | Open water (lake/pond) | -119.168659 | 47.85426 |
| P8 | Dry shrubland/rangeland | -119.688928 | 47.478636 |
| P9 | Snow | -121.103223 | 48.109005 |
| P10 | Urban | -120.508112 | 46.598891 |
Step 4.2 — Create a Spectral Profile Chart
- In the Contents Pane, right-click your 6-band composite layer
- Select Create Chart → Spectral Profile
- A Chart Properties pane will open on the right, and an empty Spectral Profile panel will appear at the bottom of the screen
- Confirm the Layer dropdown in Chart Properties shows your composite
- Under Plot Type, select Mean Line
Step 4.3 — Sample Your Cover Type Locations
You will click directly on the map to collect spectral signatures at each of your 10 sample point locations. Use the lab01_sample_points shapefile as a visual guide — click as close to each point marker as possible.
- In the Chart Properties pane, find the Define an area of interest section — you will see a row of drawing tool icons
- Click the Point tool (circle icon, far left of the drawing tools)
- Click on the map at each of your 10 sample point locations. Each click adds a new entry to the Spectral Profiles list in the Chart Properties pane
- After adding each point, double-click its name in the list to rename it to its cover type (e.g., “Forest”, “Burnt”, “Irrigated”, etc.)
- Click the colored square next to each entry to assign an intuitive color
What to expect in your chart
The figure below shows an instructor-generated example of what your completed spectral profile chart should look like. Your curves will differ slightly depending on exactly where you clicked, but the overall shape of each cover type should match these general patterns.
| Composite Band | Landsat Band | Wavelength | Spectral Region |
|---|---|---|---|
| 1 (B2) | Blue | 0.48 µm | Visible |
| 2 (B3) | Green | 0.56 µm | Visible |
| 3 (B4) | Red | 0.65 µm | Visible |
| 4 (B5) | NIR | 0.86 µm | Near-Infrared |
| 5 (B6) | SWIR1 | 1.61 µm | Shortwave Infrared |
| 6 (B7) | SWIR2 | 2.20 µm | Shortwave Infrared |
General patterns to look for across your 10 points:
- Dense forest (P1, P2): Strong rise from Red to NIR (bands 3→4), the classic vegetation “hockey stick”
- High-severity burn (P3, P4): Flat and low across all bands — little reflectance anywhere
- Dry/fallow agriculture (P5): Moderate, relatively flat — low NIR, higher SWIR
- Irrigated agriculture (P6): Similar shape to forest but lower amplitude
- Open water (P7): Drops steeply after visible bands — nearly zero NIR and SWIR
- Dry shrubland (P8): Some NIR signal but much less than forest — between burn and healthy veg
- Snow (P9): Very high reflectance across visible AND NIR bands — drops sharply only in SWIR (bands 5–6)
- Urban (P10): Moderate and relatively flat across all bands — no strong NIR peak
📸 Screenshot 4: Take a screenshot of the Spectral Signature Viewer showing the spectral curves for all 10 sample points. Make sure the legend (point labels or colors) is visible. Include this in your submission.
❓ Question 5 (Interpretation): Using your Spectral Signature Viewer, answer the following:
Which point(s) show the highest NIR reflectance (Band 4)? What land cover does this represent, and why does this surface have such high NIR reflectance? (Reference Campbell Ch. 2.)
Which point(s) show the lowest NIR reflectance? What does this tell you about those surfaces?
Snow (P9) has very high reflectance in both visible AND NIR bands, but drops sharply in SWIR (Bands 5–6). Healthy vegetation also has high NIR reflectance. Based on your chart, which band(s) best separate snow from healthy vegetation, and why might this matter for remote sensing applications in mountainous or high-latitude environments?
Urban (P10) and dry shrubland (P8) may look similar in natural color. Do their spectral profiles help separate them? Which bands are most diagnostic?
❓ Question 6 (Interpretation): Campbell Ch. 2 introduces the idea of a spectral signature — the concept that different surface types produce distinctive patterns of reflectance across wavelengths. Looking at your spectral profile chart, do the 10 sample points produce clearly distinct signatures, or do some cover types have similar spectral patterns?
Identify one pair of cover types that would be difficult to distinguish using only the visible bands (Bands 1–3), and explain why adding NIR and SWIR bands helps separate them.
Part 5 — Map Layout and Export
The final task is to produce a cartographically complete map of the Pearl Hill burn scar. A cartographically complete map is one that another person can read and interpret without your help — it carries everything they need to understand what they are looking at, where it is, and at what scale.
Your map does not need to look like the instructor example below. It should reflect your own cartographic judgment about how to best communicate this particular scene. The instructor example is provided so you can see what the elements look like in practice, not as a template to replicate.
Step 5.1 — Create a New Layout
- In the ribbon, go to Insert tab → New Layout (click the dropdown arrow below “New Layout”)
- Under ANSI — Landscape, select Letter (8.5 × 11 in)
- A new Layout tab will open alongside your Map tab
Step 5.2 — Add the Map Frame and Set Your Extent
- With the Layout tab active, go to Insert tab → Map Frame
- Click and drag on the white canvas to draw a rectangle for your map
- Double-click inside the frame to activate it, then pan and zoom to center on and highlight the burn scar — this is your subject, and it should dominate the frame
- Click outside the frame to deactivate when you are satisfied with the extent
Step 5.3 — Confirm the Band Combination
Make sure your map is displaying the NIR false color composite (composite bands 4-3-2). The layout reflects whatever is currently displayed in the map view — switch to the Map tab to adjust symbology if needed.
Step 5.4 — Required Map Elements
Every map you produce in this course must include the elements below. For each one, the rationale is explained so you understand why it is required — use that understanding to make your own decisions about placement, size, and style.
Title
Why: A map without a title forces the reader to guess what they are looking at. The title should answer: what is this a map of, where, with what data, and when?
How: Insert tab → Text. Click anywhere on the canvas and type. To edit after placing: double-click the text element. To change font size or style: select the element → Format tab → Text Symbol.
Your choice: Write a title that accurately describes this map. Include the event name, location, sensor, band combination, and date. Make it large enough to read clearly. Place it where it does not compete with the map content.
North Arrow
Why: Establishes map orientation for the reader. Essential whenever north is not self-evident.
How: Insert tab → North Arrow. A style gallery appears — choose any style that reads clearly. Click to place, drag corner handles to resize.
Your choice: Size and placement are yours to decide. The convention is unobtrusive — a north arrow that competes with the map data is too large.
Scale Bar
Why: Allows the reader to estimate real-world distances from the map. A map without a scale bar is dimensionless.
How: Insert tab → Scale Bar. Choose a style. Double-click after placing to open properties and set division units and values.
Your choice: Choose a scale bar span and interval that is appropriate for your map extent — the right answer depends on how far you are zoomed in. The one non-negotiable: scale bars must be in kilometers (SI units), not miles. Remote sensing and geographic science use metric units universally. If the scale bar shows “unknown units,” check your coordinate system (UTM Zone 10N).
Coordinate Grid
Why: Provides geographic reference so the reader knows exactly where on Earth this map is located.
How: Double-click the map frame to activate it → right-click → Properties → Grids tab → Add Grid → Graticule. Expand the graticule settings to set the interval and label options.
Your choice: Choose a grid interval that produces readable, well-spaced lines for your map extent. Labels should be legible but small enough not to clutter the map.
Band Combination Key
Why: Unlike a standard photograph, a false color composite requires explanation — the reader cannot intuit what the colors mean. A key tells them which bands are displayed and what the major color classes represent.
How: Insert tab → Text. Type a brief key directly on the layout — there is no automated tool for this since the composite has no classified symbology.
Your choice: What does the reader need to know to interpret the colors in your specific map? At minimum: which bands are shown (e.g., NIR–Red–Green as R–G–B) and what the dominant color patterns mean (e.g., bright red = healthy vegetation).
Data Source Credit
Why: Scientific and professional maps always cite their data source. This allows the reader to verify, reproduce, or extend the work.
How: Insert tab → Text. Place it in a margin area — typically small text, unobtrusive.
Your choice: Include the sensor, product level, and source at minimum (e.g., “USGS Landsat 8 OLI Level-2 Surface Reflectance, EarthExplorer”).
Author Name and Date
Why: Establishes accountability and documents when the map was made.
How: Insert tab → Text. Place in any margin.
Step 5.5 — Review Before Exporting
Before exporting, step back and ask: does this map clearly communicate the Pearl Hill burn scar? A reader who knows nothing about this fire should be able to find the burn scar, understand what the colors mean, and know where they are geographically.
Common issues to check:
- Does the burn scar actually fill a meaningful portion of the frame, or is it a small feature lost in a large scene?
- Can all text (title, labels, scale bar units) be read at normal viewing distance?
- Does the band combination key explain what the reader is actually seeing?
- Is the scale bar in kilometers?
Step 5.6 — Export to PDF
- Share tab → Export Layout
- Set File Type to PDF
- Set Resolution to 300 dpi
- Name the file
Lab01_Map_YourLastName.pdf - Export to your Lab01 folder
Standard Operating Procedure (SOP)
Complete this section in your own words as you work through the lab. Your SOP is a replicable procedure document — someone else should be able to follow it to reproduce your exact result. Write it as if addressing a technically competent colleague who has never done this specific task.
1. Objective
In 2–4 sentences, state the scientific purpose of this analysis. What environmental question are you investigating, and what does the Landsat 8 imagery allow you to observe that other methods would not?
2. Data Inventory
List each data source used, following this format:
| Dataset | Source | Acquisition Date | Spatial Resolution | Spectral Bands Used |
|---|---|---|---|---|
| Landsat 8 or 9 OLI L2 SR | USGS EarthExplorer | [your date] | 30 m | B2, B3, B4, B5, B6, B7 |
| Sample Points (lab01_sample_points) | Instructor-provided | 2026 | N/A | N/A |
3. SOP — The Workflow
Write a numbered, step-by-step account of your workflow. This should be detailed enough that another student could reproduce your analysis exactly.
At minimum, cover: (1) EarthExplorer scene search and filter criteria, (2) how you assembled the composite, (3) how you set each band combination, (4) how you used the Spectral Profile chart.
4. Results and Visual Evidence
In 3–5 sentences, summarize the key visual and spectral patterns you observed. Which band combination was most useful for identifying the burn scar? What did the spectral profiles reveal about the difference between burned and unburned surfaces?
5. Analyst’s Note
Identify at least two limitations of this analysis. These should be methodological or data-related limitations — not personal difficulties with the software.
Examples of the kinds of limitations to consider (do not simply copy these; write your own): - Temporal gap between fire event and image acquisition - Spatial resolution constraints and mixed pixels - Atmospheric correction assumptions in Level-2 data - Single-date analysis versus temporal comparison
Submission Checklist
Compile the following into a single PDF or Word document and submit to Canvas before the next class session:
ESCI 442/542 — Introduction to Remote Sensing | Western Washington University | Last updated: April 2026