We live, as humans ever have, in a complex world. The natural and human worlds today have intersected to such an extent that there are possibly no parts of the planet where the influence of humans is not felt at some level. The great patterns of change unfolding in our world influence and impact in ways and at scales that are profound. Population growth, the migration of humans into the cities, urbanisation on a global scale, the consequences of the changing climate on all life, the patterns of change affecting land and its capacity to support ecosystems, and the relationship between the ice shelves, the oceans and the atmosphere.
This web of interconnectivity makes it practically impossible for any person to fully embrace, comprehend and act upon all these complexities. Scientists seeking to develop an understanding of the world, trying to identify the causes of change, to consider potential options for the future and to find ways to advance society therefore have to develop appropriate computational tools to aid their work.
We refer to the ‘Digital Environment‘ therefore as a way of describing how we may represent our world, in digital form, through a combination of computer systems, with measurements taken from a plethora of environmental sensors, using advanced analytical and statistical techniques able to draw across vast bodies of environmental data, together informing our understanding of how these complexities impact on our lives and on society at large – approaches that demand a nexus of the fields of environmental science, of computer science, of data science and of decision and social science.
It is often said that great innovations can arise in those areas that sit between traditional disciplines, where ideas from different sources can collide with unforseen results. Constructing a ‘digital environment’ offers us great challenges, but also great opportunities, where truly interdisciplinary thinking and approaches can yield great advances in our understanding of the world. The tools that we seek to bring to bear must have the ability to capture and comprehend these complexities, to handle and assess uncertainties in an uncertain world, and to manage and manipulate a vast and broad spectrum of sources of information. For these reasons, the field of Artificial Intelligence (or AI), and its many branches, is seen to hold great promise to environmental science.
There are many forms of AI offering this promise, but one of these in particular is gaining in prominence. Machine Learning is an exciting branch of AI that is increasingly being applied in environmental science digital solutions. Machine Learning techniques allow us to process vast bodies of data originating from many sources, and to look for patterns in the data, developing deep understanding of it, and learning from these patterns to gain the ability to make predictions of given future outcomes based upon these. Of course, using computers as an aid to decision making is not a new thing, indeed computers have been used to support decisions since their very advent. The recent 50th anniversary of the lunar landings remind us of the critical role those early computers played in landing the first astronauts safely on the moon.
Today, environmental scientists use powerful high performance computers, such as JASMIN to seek to understand the complexities of our Earth and its many interconnected and interdependent systems. That complexity is itself a challenge that the modern field of Artificial Intelligence (AI) can be set to address. It has been noted that many of the statistical techniques embodied in AI approaches are not new – for example, Gauss’ work described the distributions of values in datasets and the trends apparent thereby back in the early nineteenth century. It is perhaps however the scale of the computational challenge that today demands approaches that can work with the vast lakes of ‘big data’ and associated metadata, arising from observational science, from sensor readings, from computer models, from citizen science and social media to name a few sources – data of considerable variety that arrives at high velocity, in great volumes, the three ‘V’s of big data.
Together with the new generation of analytical techniques comes the need to convey the outcomes in an appropriate and comprehensible way to a range of end users and stakeholders – from scientists, to policy makers, and from business and industrial concerns to the general public. In parallel then with the data processing, modelling and AI approaches must also come a new generation of visualisation and representation of the outcomes of these techniques. The use of intelligent mapping, of dynamic infographics, of virtual and augmented reality, of dashboards – able to portray the complexities in compelling and comprehensible ways. It is perhaps this last frame that will characterise the likelihood of the future success of the digital environment. The means to communicate and explain, and to show options and consequences and to support exploratory scenario building in forecasting the future, hind casting the past and nowcasting current events. In this way the ‘digital environment’ must embrace an arc of scientific, humanities and complex systems thinking – from data collection, to information handling, to knowledge creation.
The ‘Creating a Digital Environment’ programme described in this website is a strategic priority initiative of the UK Research and Innovation funding body, the UKRI, and within this in particular the interests of the Natural Environment Research Council (NERC) and the Engineering and Physical Sciences Research Council (EPSRC), together with strong input from the Department for Environment, Food & Rural Affairs (Defra), and its member organisations such as Environment Agency, the Joint Nature Conservation Committee, Natural England, and the Animal and Plant Health Agency. The work is aligned closely with key policy initiatives, such as Defra ‘25 Year Plan‘ for the environment. The Digital Environment programme will support leading scientific research in this field, will draw upon international expertise, and will support the creation of a community of practice for environmental scientists seeking to develop and benefit from the latest generation of tools and techniques. This website will provide a reference point for keeping in touch with this as it develops over the coming years.