In this long-awaited fourth and final installment of our education series, we turn to index layers! Remote sensing indices build on everything we’ve learned thus far by giving us access to information outside of the visible light portion of the electromagnetic spectrum, similar to false-color imagery. What makes index layers unique, though, is that they use multiple spectral bands to calculate an index value to represent a specific on-the-ground phenomenon (like vegetation health, amount of water present, etc). This allows us to both view that phenomenon represented by the index spatially, and to easily create a time-series for better illustrating change over time.
There are a number of different index layers in Lens; this piece will focus on 3 indices to give an overview of what indices are along with examples of how those indices are often used. For the latest list of index layers available in Lens, visit our help article on index layers.
NDVI - Normalized Difference Vegetation Index
NDVI measures vegetation. Throughout this article - and when exploring remote sensing indices in general - you may see the term 'normalized difference' a lot. This simply refers to the structure of the equation used to calculate this index. In the case of NDVI, the index is calculated with the near-infrared (NIR) and red bands using the equation:
NDVI = (NIR-RED)/(NIR+RED)
Healthy plants absorb red light and reflect near-infrared, meaning they will have high values for NIR and low values for RED resulting in values closer to 1 from this equation. Conversely, areas of stressed vegetation or no vegetation will reflect more red light and absorb more near-infrared resulting in lower NDVI values. Theoretically, NDVI values can range from -1 to 1, however negative values are pretty rare and almost always represent water1. So in Lens, we clip these values from 0 to 1.
Values close to 0, shown in Lens in white and red, represent bare ground or urban areas. Low and middle values, shown in Lens in yellow and light green, represent stressed or sparse vegetation. And high values near 1, shown in Lens in dark green, represent dense healthy vegetation.
Index Layer Quick Tip #1: While looking at the exact values of the indices can be useful, usually viewing changes over time is where more valuable insights come from and it’s typically not recommended to compare values of indices across different sensors/satellites. So if you want to examine vegetation on a property over time, you should stick with one sensor (like S2).
This index is certainly one of the most commonly used indices in remote sensing, but is not without limitations. One of the main limitations of NDVI is that it can saturate quickly, so in areas of dense vegetation it can be hard to distinguish which crop field or which area of a rainforest has the most dense and healthy vegetation3. Similarly, NDVI will not really provide clues around type of vegetation: a field of crops and forest can both register similarly high NDVI values. Despite these limitations, there are still many valuable use cases for NDVI.
NDVI in Action
There are seemingly endless use cases for NDVI - let’s take a brief tour through some of these as shown in Lens!
NDVI is commonly used in agricultural applications. Vegetation is monitored to see which fields are active and pinpoint the time or harvest. NDVI can also be used to identify cover cropping.
Vegetation is often monitored to pinpoint when a disturbance occurred and to watch revegetation post disturbance. Two common instances of this are wildfires and logging.
We can also use NDVI to track changes in ecosystems over time like this instance of mangrove expansion.
NDWI - Normalized Difference Water Index
NDWI provides a measure of the amount of water or moisture present. In Lens, we use this index and clip the values to show two separate layers: surface water and surface moisture. For both layers in Lens the same equation is used calculating from the near-infrared and green bands:
NDWI = (GREEN-NIR)/(GREEN+NIR)
Water has very low reflectance in the near-infrared portion of the electromagnetic spectrum, so areas of water will have higher values than dry areas.
The surface moisture layer in Lens uses NDWI to show a spectrum from dry to saturated soil and water; values are clipped from -0.6 to 0 where lower values shown in whites and tans represent dry ground and higher values shown in greens and blues represent saturated soil or presence of water.
For the surface water layer in Lens, values are clipped to -0.1 to 0.4, where low values shown in grays and tans represent dry land and high values shown in dark blue represent the presence of water. Because of the way values are clipped for the surface water layer, this layer can be used more as a binary: showing where water is present (think rivers, ponds, etc.) and where there is no standing water present.
Surface moisture and surface water, like many other index layers, can be distorted by shadows. It is always a good idea to use Compare Mode in Lens to compare the index you’re interested in alongside a truecolor image to contextualize what you’re seeing. In mountainous areas and places with complex topography, this can be a critical step to ensuring that shadows from the landscape aren’t interfering with the index values. Similarly, in areas with dense tree coverage or other vegetation blocking a top-down view of the surface of the earth, surface water and surface moisture inaccurately show a lack of water. In these cases, the vegetation layer can be more aptly used to approximate moisture: areas of dark green indicate healthy vegetation not experiencing water stress.
Index Layer Quick Tip #2: Use flexible compare mode to contextualize the index layer with true-color imagery. Think about how the topography of the area (read: shadows from terrain) could be affecting the index values.
NDWI in Action
Surface moisture and surface water have many use cases: from planning the best time of year to conduct a field visit to avoid the mud season, to monitoring flooding or restoration progress, let’s take a closer look at some examples from Lens below.
Surface moisture is a great tool to use for planning field visits to avoid the wettest times of year! This example shows an area that is consistently wet in late winter and early spring, but tends to dry out by May.
And this example shows how surface moisture can be used to monitor restoration efforts.
Surface water can be invaluable for monitoring water levels, like in this example showing a reservoir’s varying amounts of water from year to year.
Analyzing surface water is also a great way to assess the frequency of flooding. Here, we can see a river floodplain that has flooded a handful of times in the last 5 years.
NBR - Normalized Burn Ratio
NBR, also sometimes referred to as the Burn Index, measures burn severity and helps identify areas that have recently burned. Calculated using the near infrared and short wave infrared the NBR equation is4:
NBR = (NIR-SWIR)/(NIR+SWIR)
Burned areas reflect highly in the shortwave infrared portion of the spectrum and reflect low amounts of near-infrared; this is the opposite of healthy vegetation. Values in Lens range from -0.1 to 0.4, where lower values shown in black denote healthy vegetation and higher values shown in yellow represent recently burned areas. NBR values in Lens are clipped from -0.1 to 0.4 to make it easy to identify burned areas.
It is important to keep in mind that index values can vary based on the landscape and ecosystem. For example, an arid grassland in the middle of summer can show higher burn index values than a forested area experiencing no water stress. Therefore, it is important to contextualize this index over multiple years to look for unseasonal spikes in the NBR values.
Index Layer Quick Tip #3: Consider the whole time series of an index on a property within the context of the ecosystem. Are cyclical seasonal patterns expected? What might an unseasonal spike or drop indicate?
NBR in Action
NBR is commonly used for monitoring fires. This index can help pinpoint the timing and intensity of a fire.
Additionally, this index can be used in conjunction with NDVI to determine if a drop in vegetation is due to fire activity or other activity like logging or pests. In this example, we can confirm that the drop in vegetation was due to fire.
Aaaand that is a wrap on our first remote sensing educational series! From learning some of the basics of remote sensing to gaining a deeper understanding of the electromagnetic spectrum, false color imagery, and index layers, we hope this series has equipped you with more knowledge to dive deeper into Lens and expand your perceptual world. Let us know what you’d like to learn more about next!