AI in environmental protection – an overview of functionality and application scenarios

Sustainability and environmental protection are topics that are now part of our daily lives. What we eat, where we travel, the clothes we wear – each of these have an impact on our personal environmental footprint, and now the importance of these issues has made its way into the political arena. You’ve probably heard of the United Nations’ 17 Sustainable Development Goals (SDGs). These goals form a framework for achieving the targets we as a global community wish to achieve by 2030. The 17 goals address various areas of sustainability. In its new coalition agreement, Germany’s current federal government also refers to the 17 SDGs as the guiding principle of our policy and intends to closely align its actions with them. It has set out to increase the binding nature of the strategies, goals and programmes for sustainability when taking concrete government action and drafting laws.

Current environmental problems and use of artificial intelligence

We’ve been experiencing some of the direct consequences of climate change in Germany for quite a while now – including the environmental changes that occurred as a result. Increased numbers of forest fires, prolonged droughts or frequent heavy rainfall usually have devastating humanitarian, environmental and economic consequences. Pest infestations and the increase in air and soil pollutants also have a negative impact on various ecosystems, such as forests or agricultural land.

If we want to truly understand environmental problems, systematically analyse them and take action against them using IT systems, we must draw on our most crucial tool: environmental data. For the most part, this exists in the form of vast amounts of data that are rarely accessible to the public. In the best case, environmental data should be both structured and harmonised. Only then will it be possible to intelligently link this data and analyse it. AI comes into play when dealing with very large amounts of data that have to be analysed in a short period of time. AI-based systems are able to quickly and efficiently determine important environmental parameters, such as the health of leaves in canopies or the level of moisture in soil and leaves in forest areas. Various predictions can be made by intelligently linking parameters such as these with other specific data sets, such as current and historical meteorological data. AI-based systems are then able to use all the data to detect patterns, which are used as the foundation to make predictions. Forecasts such as these can form an important basis for making decisions, not only for public institutions, such as forestry offices, but also for entities that privately manage environmental resources.

In addition to the previously mentioned environmental data, both historical and current data on precipitation, temperature and soil permeability, as well as geological data on soil, can be useful and potentially interesting and relevant for various use cases. It’s also important here to have a reasonable data basis, which, in the best case, is standardised. This basis defines the quality of the data, how it’s structured and what content is available. In public administration, we deal with very large amounts of data, most of which have not been processed in an AI-friendly manner. The data must be processed accordingly and made available by the respective ministries and institutes.

But how exactly does AI process the data? When a certain amount of environmental data is available, it gets merged into a single data set. Things such as machine learning methods can then be applied to analyse and evaluate the existing data. Predictive and prescriptive analytics are key methods in this context and have already been tried and tested. By using these technical processes, you can transform complex sets of data into easily comprehensible recommendations for action. Such concrete recommendations for action can save actual lives, for example, in the field of disaster management.

Project funding by the BMUV

The German Federal Ministry for the Environment, Nature Conservation, Nuclear Safety and Consumer Protection (Bundesministerium für Umwelt, Naturschutz, nukleare Sicherheit und Verbraucherschutz, BMUV) has been promoting what are known as the AI lighthouses since 2019. These are federally funded IT projects that pursue the goal of contributing to environmental protection through the targeted use of artificial intelligence. The first of two lines of funding for the initiative has provided a total of €28 million to 28 projects. The second line of funding started in autumn of 2021. The AI lighthouses are intended to make Germany and Europe a prime location for AI technologies. They’re intended to make it clear which environmental challenges can be met using AI solutions and make a significant contribution to implementing the BMUV’s five-point programme Artificial Intelligence for the Environment and Climate.

Conclusion

Sustainability and environmental protection are topics that are now part of our daily lives. What we eat, where we travel, the clothes we wear – each of these have an impact on our personal environmental footprint, and now the importance of these issues has made its way into the political arena. You’ve probably heard of the United Nations’ 17 Sustainable Development Goals (SDGs). These goals form a framework for achieving the targets we as a global community wish to achieve by 2030. The 17 goals address various areas of sustainability. In its new coalition agreement, Germany’s current federal government also refers to the 17 SDGs as the guiding principle of our policy and intends to closely align its actions with them. It has set out to increase the binding nature of the strategies, goals and programmes for sustainability when taking concrete government action and drafting laws.

Current environmental problems and use of artificial intelligence

We’ve been experiencing some of the direct consequences of climate change in Germany for quite a while now – including the environmental changes that occurred as a result. Increased numbers of forest fires, prolonged droughts or frequent heavy rainfall usually have devastating humanitarian, environmental and economic consequences. Pest infestations and the increase in air and soil pollutants also have a negative impact on various ecosystems, such as forests or agricultural land.

If we want to truly understand environmental problems, systematically analyse them and take action against them using IT systems, we must draw on our most crucial tool: environmental data. For the most part, this exists in the form of vast amounts of data that are rarely accessible to the public. In the best case, environmental data should be both structured and harmonised. Only then will it be possible to intelligently link this data and analyse it. AI comes into play when dealing with very large amounts of data that have to be analysed in a short period of time. AI-based systems are able to quickly and efficiently determine important environmental parameters, such as the health of leaves in canopies or the level of moisture in soil and leaves in forest areas. Various predictions can be made by intelligently linking parameters such as these with other specific data sets, such as current and historical meteorological data. AI-based systems are then able to use all the data to detect patterns, which are used as the foundation to make predictions. Forecasts such as these can form an important basis for making decisions, not only for public institutions, such as forestry offices, but also for entities that privately manage environmental resources.

In addition to the previously mentioned environmental data, both historical and current data on precipitation, temperature and soil permeability, as well as geological data on soil, can be useful and potentially interesting and relevant for various use cases. It’s also important here to have a reasonable data basis, which, in the best case, is standardised. This basis defines the quality of the data, how it’s structured and what content is available. In public administration, we deal with very large amounts of data, most of which have not been processed in an AI-friendly manner. The data must be processed accordingly and made available by the respective ministries and institutes.

But how exactly does AI process the data? When a certain amount of environmental data is available, it gets merged into a single data set. Things such as machine learning methods can then be applied to analyse and evaluate the existing data. Predictive and prescriptive analytics are key methods in this context and have already been tried and tested. By using these technical processes, you can transform complex sets of data into easily comprehensible recommendations for action. Such concrete recommendations for action can save actual lives, for example, in the field of disaster management.

Project funding by the BMUV

The German Federal Ministry for the Environment, Nature Conservation, Nuclear Safety and Consumer Protection (Bundesministerium für Umwelt, Naturschutz, nukleare Sicherheit und Verbraucherschutz, BMUV) has been promoting what are known as the AI lighthouses since 2019. These are federally funded IT projects that pursue the goal of contributing to environmental protection through the targeted use of artificial intelligence. The first of two lines of funding for the initiative has provided a total of €28 million to 28 projects. The second line of funding started in autumn of 2021. The AI lighthouses are intended to make Germany and Europe a prime location for AI technologies. They’re intended to make it clear which environmental challenges can be met using AI solutions and make a significant contribution to implementing the BMUV’s five-point programme Artificial Intelligence for the Environment and Climate.

Conclusion

Drawing our conclusions from the AI projects that are currently underway will be important for the future. The experience we gain in the projects will help us to optimise AI systems that are already in operation and to develop new use cases in the environmental sector that we’ll work on in new AI projects. One challenge we’re facing that shouldn’t be ignored is the significant amount of energy it takes to run AI systems. We mustn’t forget that we’re using AI to protect the environment. It is absolutely necessary that we operate the AI systems in a way that conserves resources and uses energy efficiently. I hope that the German federal government implements the plans set out in the coalition agreement and continues to support the necessary projects.

You will find more about the exciting topics from the world of adesso in our latest blog posts.

Drawing our conclusions from the AI projects that are currently underway will be important for the future. The experience we gain in the projects will help us to optimise AI systems that are already in operation and to develop new use cases in the environmental sector that we’ll work on in new AI projects. One challenge we’re facing that shouldn’t be ignored is the significant amount of energy it takes to run AI systems. We mustn’t forget that we’re using AI to protect the environment. It is absolutely necessary that we operate the AI systems in a way that conserves resources and uses energy efficiently. I hope that the German federal government implements the plans set out in the coalition agreement and continues to support the necessary projects.

You will find more about the exciting topics from the world of adesso in our latest blog posts.