New Themes, Multi-Temporal Scripting and Other Improvements in EO Browser
New Education Mode, Multi-temporal scripting, Better control of time intervals and six other cool features at the end of your fingertips!
1. Education Mode
A new Education mode was added to EO Browser. You can toggle between the Normal and Education mode by clicking on the little graduation cap icon on the top right:
Education mode is a powerful tool for showcasing the usefulness of satellite data for various industries and phenomena, especially to students.
Currently, there are 12 themes, which you can see on the left.
When you select a theme, it’s best to first look at the prepared pins and see some areas of interest. After selecting one, you can test the available visualization layers. You can also search by the available datasets, as is the default in EO Browser.
The best-fitting visualization layers were chosen for each theme and all have descriptions with references. Moreover, 157 unique pins were added. The pins were chosen to either showcase the diversity of the phenomena or the usefulness of the composites and custom scripts for detecting, for example, urban areas, burnt areas, chlorophyll content, vegetation health, etc.
Additionally, all the visualization layers, that return one value (such as indices) have statistical information service set up.
Atmospheric correction is done by default in Education mode; the L2A dataset is always on top and preselected for the Sentinel-2 satellite and all the pins, where it was available, use the L2A data.
Below we have selected a couple of example pins from different Education themes. To find out more about them, click on the links in the descriptions and visit EO Browser Education, where you can explore so much more.
2. Advanced Control of Time Intervals
In EO Browser, you can now select Timespan on the right to bring up the advanced acquisition time editor, where you can manipulate the time range of data shown, beyond the default “daily” interval.
The new feature makes it possible to narrow down the results if we know the image we need was captured at a specific time range.
This is useful for the selection of specific data orbits when they overlap or when one would need to select either ascending or descending Sentinel-1 orbits, and for multitemporal processing (see next chapter).
In the image below (left) you can see two overlapping Sentinel-3 OLCI orbits on the same day. Let’s suppose we only want to keep the earlier acquisition orbit on the right. With the new tool, we can do so by setting the acquisition time from 00:00 to 10:30 AM (as is shown in the image above), which selects only the earlier orbit. Should we want the later acquisition orbit, we should set the acquisition time to be 10:30 to 23:59 instead.
3. Multitemporal Scripting
Multitemporal scripts now work in EO Browser!
To make it work, you should:
- Convert your script to Version 3
- Add mosaicking: Mosaicking.ORBIT to your function setup
- Set up your time range in the new timespan feature (see 2. Advanced Control of Time Intervals). Setting up time ranges is much easier and you don’t even need the filterScenes function in most cases.
To convert your multitemporal script, which already works in Playground, to Version 3, you need to add //VERSION=3 at the top and edit the function setup(), which is structured a bit differently than you might be used to:
The following is a complete example of the Harel Dan’s multitemporal NDVI script applied to the agricultural fields of Saudi Arabia, which are also one of the Agriculture theme pins.
A timespan was set up from the beginning of April to the end of June 2019:
The script calculates the average NDVI values for each of the three months and then applies them to the RGB channels. Red fields in the image above had high NDVI values in the first month (April) and low NDVI values in the other two. Pink fields had high NDVI values in the first (the red channel) as well as the third (blue channel) month. Check it out in EO Browser.
Note that processing large amounts of data takes longer so you will need to wait a bit for the results.
4. Better Timelapse Filters
When creating a timelapse, you now have a much finer data search control. Aside from previously available cloud coverage and monthly selection filtering, you can also select whether you want orbit, daily, weekly, monthly or yearly data. This way, you can for example only return one image per month for the last couple of years, or one result per week for the last 2 months. As EO Browser timelapse search only allows for 300 results, this control makes it possible to create time-lapses over longer time intervals.
In the timelapse below we can observe weekly changes of CO concentrations above Ghana, Togo, and Benin from April 2018 to March 2020. We can see that the concentrations rise in the winter months of 2018 and 2019. This is due to wood and coal burning for heat. As we have selected one result per week, we were able to get weekly results further back in time and confirm that CO concentrations rose consistently in winter of both 2018 and 2019.
To make use of the new feature even more, we could only select the November-February months and observe the increasing and decreasing of CO concentrations on a daily basis.
5. Borders in Timelapse
When creating time-lapses, you can now enable the borders checkbox and country borders will be displayed. This is useful as with some datasets, such as Sentinel-5P, it can be difficult to know which area you are looking at.
On the timelapse below, you can observe the vegetation changes during the 2018 drought in Luxembourg (with the Barren soil script). The red patches of land indicate barren soil and green indicates vegetation. In August, you can observe the sudden disappearance of vegetation from lush green to barren red. This is due to drought, as vegetation should be at its peak in August. In September, as the drought ends, you see the greening of the vegetation again.
Thanks to the new borders feature, you can see the outline of Luxembourg and its regions, making it clearer which area we are looking at.
6. Statistical Analysis for More Datasets
Sentinel Hub now supports statistical info service for Sentinel-3, Sentinel-5P, and Landsat. In EO Browser, statistical info service is already set up for you to use on Sentinel-3 OLCI OTCI, Sentinel-5P data products, and Landsat NDVI and Thermal layers.
You can also set up statistical info service for your own layers in the Sentinel Hub Configurator by creating a new layer called __FIS_ID, where ID is the same as your layer ID. Your statistical info service layer should also only return one value. For example, if your NDVI layer ID is NDVI-1, your statistical info service layer should have both ID and name called __FIS_NDVI-1, returning the NDVI values instead of the usual colored visualization in the evalscript (i.e. return[ndvi]). If you set it up correctly, you will be able to use the statistical information on your layer, displayed in EO Browser.
In the images above you can see Landsat 8 NDVI for March 2020 (left) and August 2019 (right). A point pin, as shown in the images, was chosen for the statistical information below. As is clear from the images, seasonal vegetation changes between March and August are significant, which can be confirmed by inspecting the following statistical info service results.
7. Layer Descriptions Set Up
Layer descriptions with references have been set up for most layers in EO Browser, including all default Sentinel-2 layers. In time, we plan to add the descriptions to all the available EO Browser layers. In Education mode, descriptions are already set up for all layers.
8. Sentinel-3 SLSTR Band Info Added
As with the other datasets, Sentinel-3 SLSTR custom bands now display information on mouse hover, to make it easier to select the most suitable ones for your composites. They are separated into two groups, Reflectance, and Brightness temperature.
9. MODIS Visualizations Added
If you need global daily coverage, MODIS, with a resolution of either 250, 500 or 1000 m, is free to use in EO Browser. The data is available since February 2000, so you can examine data 20 years to the past. We have expanded the visualizations, so now there are 5 of them: true color, false color, NDVI, Normalized Difference Salinity index, and the NDWI.
Do you like what you are reading and want more? Stay tuned, we are working on some cool new stuff. Subscribe to our newsletter and follow our Twitter account, so you don’t miss any awesome features.