The Winning Ideas of the Climate Change Custom Script Contest

Featuring scripts and written stories using open EO data

Authors: Monja Šebela, Sabina Dolenc

While we are on a mission to find new and innovative ideas and scripts, organizing Sentinel Hub Custom Script contests is one of our things. Together with our partners, the Euro Data Cube, Copernicus EU Earth Observation program, and the European Space Agency (ESA), we’ve been doing this already for four years in a row now. For this year’s special edition, we’ve teamed up with the World Meteorological Organization. The contest aimed to demonstrate the power of Earth Observation (EO) data in helping detect the effects of climate change. It started on March 15, 2022 and ended on September 11, 2022.

The authors behind the ideas submitted to our contest used open and free Copernicus resources via Sentinel Hub and EO Browser, a free* web application for browsing, viewing, analysing, and downloading remote sensing data. They could participate with custom scripts demonstrating efficient change detection related to climate change or with a written story focusing on how to communicate climate change via EO data.

Our jury of experts awarded the best two custom scripts and the best two written stories (the results can be found here). To find out more about all entries, keep reading this blog post.

Custom Scripts

The following scripts were created using Sentinel Hub’s custom scripts and EO Browser, and added to our custom script repository.

🥇 Detection of Lake Extent Changes

Author of the best custom script: Jan Landwehrs

The winning change detection script excels at displaying water extent changes by comparing two specific dates. It can be used to discern how much water level decreased after a drought, increased after a flood, or decreased through many years. These changes are often caused by climate change.

The script works with Landsat data, and supports data fusion, so the user can compare very long time ranges with the earlier image from e.g. Landsat 4–5 TM and the later image from Landsat 8-9. This way, we can make use of a full 50-year-long archive of Landsat data.

China’s largest freshwater lake, Poyang lake, experienced an extreme shrinkage in August 2022, associated with a severe drought and heat wave in Southern China. The result of the winning script can be seen in the third image, which shows differences between August 1988 with Landsat 4–5 TM and August 2022 with Landsat 8–9. Red color indicates water extent reduction, dark blue indicates water extent increase, and light blue shows where water was present on both dates. 🌐

Read more about the script on our Custom Script Repository.

🥈 Normalized Difference Yellowness Index (NDYI), Visualisation of Blooming in an Amazonian Tree Species and Canola Crops

Author of the script: Mario Alberto Guzmán Soza

Our silver award went to script, which uses the normalized difference yellowness index, calculated as a normalized ratio between the Sentinel-2 blue and green band, to detect the blooming of Schizolobium parahyba and canola crops. The former is a plant species common in the Amazon. Its flowering coincides with the dry season between May and July, which is closely associated with climate variations. An alteration of its phenological cycle can therefore be considered an indicator of climate change. The index was adjusted empirically to increase detection accuracy so that NDYI < (0.02–0.08) indicates blooming trees, and NDYI < (0.08–0.13) indicates flowering canola crops.

A reference image, showing all pixels in a natural color composite (left). The result of the script, where blooming canola crops are displayed bright yellow, with other areas returned in natural colors (right). This agricultural area is located in the province of Alberta, Canada (image captured on June 29, 2021). 🌐

Read more about the script in one of our guest blog posts and on Custom Script Repository.

Simple Water Bodies’ Mapping — SWBM

Author of the script: Mohor Gartner

This script from Mohor detects water areas and can be used both with Sentinel-2 and Landsat. Water detection is done based on several indices, and calibration of index thresholds is possible as well. The script can be especially helpful in water level fluctuation monitoring, coloring water areas in blue.

Left: We see flooding due to a damaged Oroville Dam (Sentinel-2 L2A, January 30, 2017 🌐 — December 21, 2017). Although the quality of the dam played a role, climate change is also to blame. When a drought was followed by extreme rainfall, dam operators wanted to store as much water as possible, causing levels to rise unexpectedly fast. Right: The floods in Pakistan (Landsat 8–9 L2, June 10, 2022 — September 6, 2022) were caused by unusually heavy monsoon rains and melting glaciers after a severe heat wave.

See the script on our Custom Script Repository.

Thermal Visualization and Water in Wetlands Script

Author of the script: Barbara de los Angeles Ortiz

The script combines water body detection with thermal data in a single script, returning blue color when surface water is detected, and a thermal visualization colored red to black when no water is detected. The author used a combination of Landsat 8–9 Level 1 and Level 2 data.

Wildfires in Argentina. Timelapse from November 2021 to March 2022 (weekly intervals). Data acquired by Landsat 8–9, USGS. 🌐

During an unprecedented heat wave and ongoing drought, Argentina suffered a series of forest fires, as temperatures in South America soared to more than 40°C. At the time, Argentina was the hottest place on the planet. Since the fires began in December, more than 520.000 hectares have burned, killing wildlife and livestock, decimating pasturelands, and destroying crops. Above, you can see a timelapse of the area from November 2021 to March 2022, showing how the humidity of the wetlands and the temperature on the surface varied.

See the script on our Custom Script Repository.

S2L2A Enhancement using S3SLTR F2 For Wildfire Detection

Author of the script: Benjamin Kuo

Benjamin’s script uses data fusion of two different satellites, Sentinel-2 L2A and Sentinel-3 SLTR, to better highlight areas burning due to wildfires. The script adjusts the brightness of the Sentinel-2 L2A true color composite by using scaled values of the Sentinel-3 SLSTR thermal F2 band, which increases the brightness of burning areas and decreases the values of other areas. The script has been tested against a small number of wildfires in California and Ukraine (image below).

Wildfire in Irpin, Ukraine with Copernicus Sentinel-2. Image acquired on March 28, 2022. 🌐

See the script on our Custom Script Repository.

Assessment of Spatio-Temporal Variations of Water Level in Lakes Using a Multitemporal Script

Author of the script: Fernando Rodr’guez Brizuela

Fernando’s script assesses spatial and temporal variations of water levels in lakes, caused by climate change. Water level measurements are especially useful for closed-basin lakes, where long-term fluctuations can be related to similar fluctuations in large-scale climate oscillations. The script is made for Landsat 5 TM and uses the Modified Normalized Difference Water Index (MNDWI) in the three RGB color channels, each corresponding to MNDWI values on a particular date.

The results of the script over the Mar Chiquita Lake, Argentina. During 2003 and 2011 (left), the lake underwent a water level decrease. Different colors correspond to MNDMI values between the three dates. Between 1990 and 1996 (right), the lake water levels initially increased (between t1 and t2), then decreased (between t2 and t3).

See the script on our Custom Script Repository.

Land Surface Temperature Comparison

Author of the script: Mohor Gartner

Mohor created this script to compare the maximum land surface temperature (LST) of two selected periods, using two instances of Landsat 8 in a data fusion script. The red output color tells us that the maximum LST was higher in the primary Landsat 8 data source (primary time frame), and blue tells us that maximum LST was higher in the secondary Landsat 8 data source (secondary time frame).

Comparison of the maximum LST over lakes Bled and Bohinj in 2013 (secondary time frame) and 2022 (primary time frame). In 2022, there was an extreme heat wave. We can see that Lake Bohinj had higher temperatures in 2022, while for Lake Bled the opposite is true; it had higher temperatures in 2013.

See the script on our Custom Script Repository.

Written Stories: How to Best Communicate Climate Change With EO Data

Seven insightful remote sensing stories were submitted to this special edition of our contest. Read the short descriptions and dive into full stories.

🥇 When it Comes to Wildfires in Siberia — Every Degree Matters

Author of the best written story: Artyom Tadzhibaev

The first place in the story category went to Artyom, who explored the devastating 2021 wildfires in Siberia. Artyom looked at the conditions that made such widespread wildfires possible, their progression and consequences. The catastrophe occurred due to a perfect cocktail of conditions; in June, the amount of precipitation was the second-lowest since 1888, with the air temperature reaching 35 degrees in the city of Yakutsk. The matter was made worse, as the temperatures didn’t drop below 30 degrees for a whole month, so only a single lighting storm was enough to ignite the inferno. By the middle of August, when the 129 fires were extinguished, at least 9200 km² of land with 31 houses burned down and at least one life was lost. Harmful smoke reached as far as Yekaterinburg 6499 km away, posing health risk to residents and lowering the global albedo, contributing to long-term temperature increase.

Wildfires around settlements on July 18, 2022 with Copernicus Sentinel-2 (left), and July 28, where we can see lots of hotspots and
enormous smoke clouds.

The author also shows how both the number of wildfires and the burned area per wildfire increased over time in Yakutia. Climate models predict the temperatures will keep increasing in the future, contributing to more common and more devastating wildfires.

Read the full story here.

🥈 Nature’s Wrath — 2022 Silchar Floods Assam

Author of the second best written story: Pawan Muddu

Pawan won second prize for his story about Silchar flooding in Assam, India. The flood occurred due to a dam breach on the Barak river, following torrential rainfall in March. The rainfall in March arrived a month earlier than expected, with expected rainfall shifting since the 1950s, causing longer dry spells with heavy rain in between. Due to climate change, extreme rainfall events are expected to increase in the future, causing more floods in this region.

As shown in this SWIR timelapse, as major embankments were breached, water Inundated into the Alluvial plains which affected major farmlands and households in the adjoining villages near the town. Large-scale erosion has been reported in Baksa, Barpeta, Biswanath, Bongaigaon, Kokrajhar, Lakhimpur, Morigaon, Nagaon, Nalbari, Sonitpur, South Salmara, Tamulpur and Tinsukia districts, while Cachar and Tinsukia witnessed urban flooding. 🌐

The Silchar flooding affected close to 5.4 million people, destroyed 548 houses and caused over 200 lives to be lost. As many as 312.085 people had to take shelter in 560 relief camps, with infrastructure like dams, bridges, and roads damaged. Large-scale erosion has been reported in several districts as well.

If you’d like to help, you can donate to the children of the flood victims in Silchar Floods. For more information, read Pawan’s blog post about his story. You can also read the original story as it was submitted to the contest here.

Sentinel-2 Imagery Quantitative Analysis: The Evolution of Surface Waters of the Po River in 2020–2021–2022

Authors of the written story: Niccolò Tolio, Carlo Masetto, Umberto Trivelloni, Silvano De Zorzi

The authors of this story look into how the 2022 drought impacted the Po river basin. 2022 was the hottest year ever for Italy, with a recorded temperature increase of + 0.76 °C compared to the historical average. Increased temperatures and low rainfall impacted many water resources, including the largest river in the country, the Po river. The authors used Sentinel-2 imagery and supervised classification to quantify the reduction in surface water extent and the increase of sand island surfaces in a selected part of the river basin. The results were alarming — water surface extent decreased as much as 35 %, and sand cover increased a whopping 280 % compared to 2020!

Supervised classification (Random Forest model) of the Po river basin. Acquired on July 22, 2020 (left) and July 22, 2022 (right)with Copernicus Sentinel-2.

Read the full story here.

Call Of The Forest

Author of the written story: Preethi Balaji

Preethi used several EO sources and scripts to analyse the area burned after the wildfire at the Bandipur National Park, India. He even modified our data fusion fire progression script to improve the final result.

Massive fires broke out in the park between February 23 and 26, 2019, destroying over 4046 hectares of forestland. Reports state that the fires flared up due to climate change and rapid growth of dry grass.

SWIR composite (left) on a date before the wildfire (February 23, 2019), and a date after the fire (February 26, 2019) 🌐. Burned area progression with the modified fire progression script (left). 🌐 Acquired by Copernicus Sentinel satellites.

Read the full story here.

The Drought of NW Iberian Peninsula — Reflection of a Water Reservoir

Author of the written story: Ana Sofia Souto

Ana used Sentinel-2 imagery to explore the effects of the 2022 drought on the area of the Alto Lindoso reservoir dam in Portugal. By inspecting the area of a polygon drawn around visible water in the reservoir, she found that the water surface area decreased from almost 8 km² in 2020 to just 3.2 km² in 2022. The scarcity of water impacted the hydroelectric power plants and affected the whole energy sector, increasing electricity prices. The drought also impacted vegetation and crop productivity, as dry soil caused significant vegetation stress and deficit.

Water extent of Alto Lindoso reservoir in August 22, 2020 (left, 🌐) and August 22, 2022 (right, 🌐). Acquired by Copernicus Sentinel-2.

Read the full story here.

Overflow Irpin River

Author of the written story: Oleksandr Gordiienko

Oleksandr used a SWIR (short wave infrared) composite to show us how the Irpin river overflew after the soviet era Irpin dam was opened to stop Russian forces from reaching Kyiv. The breach flooded the Demydiv village, which is still partially underwater 9 months later. The area is interspersed with numerous reclamation canals and was due for a massive housing construction project. The floodplain preserves one of the largest peat massifs of Kyiv Polissya and now that the waters are back for good, many hope the ecosystem will be able to recover.

Irpin river overflow before and after the dam was breached. The breach occurred on February 25, 2022. Images are acquired with Copernicus Sentinel-2. See the location after the event: 🌐

Read the full story here.

Seeing is Believing

Author of the written story: Karl Chastko

Karl wrote a story about the importance of Earth Observation tools on our perception of the globe. Astronauts who see the Earth from outer space sometimes experience the Overview shift — a cognitive shift in awareness and perception of our world, where they realize that the Earth is not just beautiful, but also fragile. Remote sensing tools can help ordinary people experience something similar without leaving the ground. Without the aid of remote sensing, it would be very difficult to communicate the spatial and temporal scale of the serious effects climate change has on our planet.

The ability to capture data over vast areas allows us to observe the impacts and consequences
of human development such as large algae blooms resulting from sediment
runoff, nutrient pollution and warming ocean temperatures. The large-scale algae blooms can lead to
oceanic dead zones, which threaten people and nature alike. This is algae bloom in the Baltic sea, captured on July 17, 2018 with Copernicus Sentinel-2. 🌐

Read the full story here.

Congratulation to the winners and thank you to all participants. For further reading we recommend the Sentinel Hub Educational page. Watch also the Custom Scripts webinar to learn more about satellite imagery and custom scripts. You can also visit a dedicated topic in the Sentinel Hub Forum for more information.

To get an impression of how easy it is to write a simple and useful script, we also invite you to read one of our previous blog posts:

Why join the next Sentinel Hub Custom Script Contest

Check out also the list of some useful information on processing images using the custom scripting in EO Browser.

Other useful links

Step into the Beautiful World of Custom Scripts

Get Creative and Solve a Remote Sensing Problem at Home from your PC and Win!

If you want to learn more about Sentinel Hub, make sure to listen the MapScaping Podcast:

*Updated on 7 August 2024:

Also read our step-by-step guide for Copernicus Browser — especially recommended for those who have been using EO Browser and want to continue with a free, public tool.