Studying grassland from space
Commonly used grassland is home to a high degree of biodiversity and fulfills an important climate-protective function as a carbon sink and also serves for animal feed and food production. However, these ecosystem services are jeopardized if productivity on these lands is maximized and their use is thereby intensified. Until now, data on the condition of pastures and pastures in Germany were not available for larger areas. in the news Remote sensing of the environmentresearchers at the Helmholtz Center for Environmental Research (UFZ) have now described how satellite data and machine learning methods allow us to assess land use intensity.
The Sentinel-2 space mission began with the launch of the Earth observation satellite Sentinel-2A in June 2015 and Sentinel-2B was launched in March 2017. Since then, these two satellites have orbited at an altitude of nearly 800 kilometers and, as part of the Copernicus program of the European Space Agency (ESA) that provides data for, for example, climate protection and land monitoring. Every three to five days, they record images in the visible and infrared ranges of the electromagnetic spectrum, which at very high resolutions up to 10 meters provide a strong basis for detecting features such as changes in vegetation.
An interdisciplinary team of researchers from the Helmholtz Center for Environmental Research (UFZ) used this publicly available data to study the land use intensity of German grasslands for the years 2017 and 2018. According to the Federal Statistical Office, these grasslands cover an area of approximately 4.7 million hectares and thus almost 30 percent of all agricultural land. “We need more information on the land use intensity of grasslands to better understand the stability and functioning of our ecosystems. The more intensively grassland is used, the greater the impact on primary production, nitrogen deposition and resilience to climate change,” said lead author Dr. Maximilian Long. He is a scientist at the UFZ Department of Remote Sensing, which is embedded in the Remote Sensing Center for Earth System Research, which was jointly established by the UFZ and the University of Leipzig.
A condition for the long-term preservation of grassland is underlying management, such as mowing or grazing. When not in use, the areas come into contact with shrubs. But the intensity of grassland management is critical to their ability to deliver ecosystem services. However, no country-wide data is available on how farmers manage their grassland. The UFZ scientist has now used the satellite data with a resolution of 20 meters to derive conclusions about mowing frequency, grazing intensity of cattle, horses, sheep and goats and fertilization in Germany.
“The magnitude of these three types of management is critical to the intensity of use,” Lange says. He defined mowing frequency classes from 0 (not mowed) to 5 (mowed five times a year) and calculated a grazing intensity from 0 to 3 (heavily grazed) from a mix of livestock numbers, species and age. For fertilization, he distinguished between fertilized and non-fertilized. He combined these three categories to derive an index indicating the management intensity of a grassland area, ranging from ‘comprehensive’ to ‘intensive’.
He used artificial intelligence (AI) to derive information about the three usage parameters based on the multidimensional data the researchers extracted from the satellite images. “AI can very efficiently derive information from data that is too complex for humans to understand. Reference data can be used to train machine learning algorithms to identify patterns in the satellite data that we can then evaluate and apply to draw conclusions for large areas. to lead.” he says.
Lange obtained the reference data from the field data of three biodiversity explorations sponsored by the German Research Foundation (DFG) in Hainich, Schorfheide and the Swabian Alb. Since 2006, various experiments have been conducted there in long-term studies on grassland with different levels of land use intensity. These experiments investigate topics such as the influence of land use on biodiversity and the effects of changes in species composition on ecosystem processes.
Lange used two algorithms to evaluate how accurately machine learning recognizes actual grassland use from the satellite data: Random Forest, a standard remote sensing method for classifying land cover, and CNN (Convolutional Neural Networks), a deep learning method primarily used in image processing. The result: “Both methods are a good representation of reality, and the CNN method is slightly better,” he says. The CNN method allowed the UFZ researcher to approximate the data from the DFG Biodiversity Explorations, which ranged from 66 to 85 percent (grazing intensity 66 percent, mowing regime 68 percent, fertilization 85 percent) for the 2018 sample. Random Forest-based results were slightly off. lower for all three parameters.
This is a high classification accuracy for comparable ecological remote sensing studies, but could be further improved if more grassland use data were available. “The more data that can be used to train a deep learning method and the more accurate this data is, the more accurate the results will be,” Lange says. In a next step, he tested the plausibility of the results in four sample regions in Germany. Two of these regions (Oberallgäu and Dithmarschen) are known for their intensive grassland use, while one near the Rhön biosphere reserve is only moderately used and the other, a nature reserve in Saxony-Anhalt, is only extensively used. This comparison also provided a good match between the remote sensing-based results and the actual data.
Overall, the UFZ team found that grassland in Germany was used less intensively in 2018 than in 2017. “This is mainly due to the drought in 2018 and the associated loss of grassland productivity,” says Dr Daniel Doktor, last author of the publication and head of the UFZ Land Cover & Dynamics Working Group. For example, the calculations show that 64 percent of the grassland was not mowed in 2018, while this value was only 36 percent in 2017.
“The results also show the differences in management in Germany. Management is often very intensive in regions such as Allgäu or Schleswig-Holstein, while it is much more extensive in Brandenburg or parts of Saxony,” he says. But this evaluation is just the beginning. More accurate management data from other regions of Germany is needed to draw even more precise conclusions with the machine learning algorithms.
Intensively managed grazing can increase profits, improve the environment
Maximilian Lange et al, Mapping Land Use Intensity of Grasslands in Germany with Machine Learning and Sentinel-2 Time Series, Remote sensing of the environment (2022). DOI: 10.116/j.rse.2022.112888
Quote: Studying Grassland from Space (2022, June 9,), retrieved June 9, 2022 from https://phys.org/news/2022-06-grassland-space.html
This document is copyrighted. Other than fair dealing for personal study or research, nothing may be reproduced without written permission. The content is provided for informational purposes only.