4 Resolutions of Remote Sensing – An Archaeological Perspective

I am fortunate to be part of two publications that are great examples of cutting-edge work using remote sensing in archaeology. To celebrate these publications, this post provides a succinct overview of a foundational concept in remote sensing while providing archaeological examples. Read on to learn more about the 4 resolutions of remote sensing from an archaeological perspective.

Background

Remote sensing is a term used to describe acquiring data from a distance. In archaeology, we use a variety of remote sensing methods. For example, ground-penetrating radar and magnetometry are two methods used to detect underground archaeological features (e.g., buried buildings or even entire settlements).

Airborne platforms like satellites are also a popular remote sensing method in archaeology. They are especially useful for obtaining information on a regional scale. Different satellites offer different types of data and data quality. Remotely sensed images are generated based on four different types of resolutions:

1. Spectral
2. Spatial
3. Temporal
4. Radiometric

The next section details the 4 resolutions of remote sensing.

4 Resolutions of Remote Sensing

Spectral Resolution

Spectral resolution is the number and size of bands in the electromagnetic spectrum that a remote sensing platform can capture. For example, the first two Landsat satellites use a multi-spectral scanner (MSS) and captured images using four spectral bands (green, red, and two near-infrared bands). On the other hand, hyperspectral platforms (e.g., Hyperion) can capture hundreds of bands on the electromagnetic spectrum.

In a recent paper spearheaded by my colleague Alexander Sivitskis, we used hyperspectral imagery to detect natural occurrences of chloritite in Oman.

You might remember an earlier blog post detailing the discovery and publication of the archaeological site of Aqir al-Shamoos in Oman. This find was important because it was the first known site in Arabia that demonstrated soft-stone production, but we did not know the raw material source location(s).

‘Soft-stone’ is a colloquial term and ‘chloritite‘ is a technical term that refers to the rock.

What does this have to do with spectral resolution?

The chloritite at Aqir al-Shamoos has a distinct spectral signature (like a fingerprint). We used hyperspectral imagery to target potential sources of chloritite and compare spectral signatures to locate potential raw material sources for the site.

In order to obtain a detailed spectral signature and detect chloritite, a high spectral resolution was crucial.

Spatial Resolution

Spatial resolution measures the smallest angular separation between two objects. For satellite images, this is represented in pixels and the spatial resolution for a given image is noted as how many meters that pixel represents. For example, the satellite SPOT 4’s multispectral scanner has a spatial resolution of 20m. This means that each individual, square pixel represents a spatial area of 400 square meters. There are instances in which pixel size and resolution are not the same, especially when multiple images are combined and the pixel sizes are averaged to represent a larger area.

In a recent paper led by my colleague Frances Wiig, we used specific bands (X- and C-) from satellite Synthetic Aperture Radar (SAR) to map a subsurface water channel at the archaeological site of ‘Uqdat al-Bakrah (Safah) in northern Oman.

The site of ‘Uqdat al-Bakrah was discovered in 2012 and yielded hundreds of Bronze objects. Along with the finds, the curious location at the edge of the Rub al-Khali sand dune desert piqued the interest of a number of archaeological teams.

The ArWHO team focuses on understanding the long-term role of water in Oman. We conducted a geophysical survey in 2017 and we were able to detect a channel-like feature. This is significant. Why? We not only learned a bit more about the paleo-environment, but we can work towards uncovering how the environment supported potential human activities at the site like charcoal production or metalworking.

What does this have to do with spatial resolution?

This study focuses on the methodology used to detect and map the channel-like feature our team discovered at ‘Uqdat al-Bakrah. The C- and X-bands have a shorter-wavelength than the commonly used L-band, BUT they have a higher spatial resolution. L-bands have the ability to penetrate more deeply into soils, but at shallow sandy sites like ‘Uqdat al-Bakrah the increased spatial resolution of C- and X-band data can help researchers to better see features.

Temporal Resolution

Temporal resolution refers to the frequency at which imagery is recorded for a particular area. For example, the MODIS satellite captures the same area every one to two days, while most Landsat satellites take images of the same area every 16 days.

Depending on the question or phenomenon being observed, temporal resolution could play an important role in imagery selection. A recent paper by Emily Hammer and colleagues harnesses both high spatial resolution and high temporal resolution to examine patterns of destruction from looting, agricultural activity, military occupation, urban growth, mining, and other types of development at 1000 archaeological sites across Afghanistan from 2001-2017. They used various types of satellite imagery including DigitalGlobe and CORONA and discerned interesting trends. For more information, check out their paper, ‘Remote sensing assessments of the archaeological heritage situation in Afghanistan‘.

Radiometric Resolution

Radiometric resolution is the sensitivity of a remote sensing platform to detect slight differences in energy, specifically, radiant flux (radiant energy emitted per unit time).

Remote sensing platforms typically use either a passive or active sensor. Passive sensors record electromagnetic radiation, which is reflected from the earth’s surface. Active sensors coat the earth’s surface in machine-made electromagnetic energy and record the quantity of radiant flux that is emitted back to the sensor.

Check out this list of active and passive sensors.

Radiometric resolution is measured in bits (a number to the exponential power of 2) and the higher the number, the finer the radiometric resolution. For example, the first Landsat satellite has a 6-bit radiometric resolution; however, Landsat 4 and 5 have a finer radiometric resolution of 8 bits.

Radiometric resolution is a bit tricky to understand and I don’t have any explicitly archaeological examples clearly demonstrating this concept (if you do, let me know), but it is important.

Closing Thoughts

Is there one resolution to rule them all? At the moment, there is no way to maximize all four resolutions and, for this reason, trade-offs must be made. However, as this post hopefully demonstrated, resolutions are often entangled with one another.

When it comes to archaeological research, resolution considerations are important when selecting imagery for a given research question or project. Understanding how these concepts work will help you choose which remotely sensed imagery fits your archaeological research goals.

Have you used remotely sensed imagery in your research? Do you know great examples highlighting these concepts? Let me know in the comments section below!