DECIDE: Delivering Enhanced Biodiversity Information with Adaptive Citizen Science and Intelligent Digital Engagements

Biodiversity is under increasing pressure, with consequent impacts on the benefits people gain from nature. This means that it is vital to include biodiversity in our decision-making and for this we need high quality, fine-resolution, spatial biodiversity information. With this information we can better value nature, and this can be done formally through a process called 'natural capital' assessment, such as by government agencies or local economic partnerships. We also need this information to develop better plans for protecting nature, undertaking ecological restoration to develop resilient ecological networks, and make good decisions about infrastructure development (to achieve net biodiversity gain, as is the ambition in Defra's 25 Year Environment Plan). Much of our existing biodiversity information comes from volunteer-collected species records (a process often called 'citizen science'). However, in many cases, people record where and when they want - leading to large spatial unevenness in recording, both at a national scale and at a local scale. The people and organisations who need to use biodiversity information don't simply require more records: they require better information. This requires us to construct good biodiversity models generated from the available data, communicate these models well, and preferentially target effort to add records from times and places that optimally improve the model outputs. This project seeks to achieve all of this by addressing three important questions. Firstly, can we enhance existing biodiversity information through near real-time, fine resolution, species distribution models? Secondly, can we make biodiversity information more accessible and useful to end users through data flows and automated data communication? Thirdly, can we encourage adaptive sampling behaviour in recorders, by using intelligent digital engagements, so that they re-deploy a portion of their effort to optimally improve biodiversity models? Our team is expertly placed to address these questions because we are a multidisciplinary team (environmental, computer, social and data scientists), and we will use a service design approach that actively engages data users (from national to local levels) and biodiversity recorders alongside the research team. In this project we will produce fine-resolution distribution models for about 1000 insect species across the UK (in this study focusing on butterflies, moths and grasshoppers) using earth observation sensor data, and a data lab (an online analysis platform) to automatically update outputs as new data are available. It is important to communicate these results and their uncertainty so, in collaboration, with data end users we will develop interactive and automatically-generated visualisations and text to do this effectively. We will also develop ways of assessing when and where new data will be most valuable in improving the model outputs. This, when combined with constraints (such as land access or people's recording preferences) will be communicated to recorders as bespoke recommendations via a web app. This will be developed for recording butterflies and grasshoppers (a sunny day activity), and recording moths (supported by our provision of portable, low cost light traps). We will engage recorders through established recording projects across the UK, including with partners in London (many people, but relatively few biodiversity data) and North and East Yorkshire (fewer people, and a wide variety of land uses). Throughout this project our work flows will be implemented in an data lab, so they will be flexible for use with any species and indeed could be adapted for any environmental data. The outcome of this project will be a process for enhancing biodiversity information that can be incorporated into existing recording projects and data streams, so that the outputs will be accessible and useful, for the benefit of nature and people.

Our project seeks to deliver improved biodiversity information for end users. At this stage of delivery the outputs will be fine-resolution species distribution models for c. 1000 species of insect. The information will be comprehensive because it will be based on model outputs, and will be fine-resolution (expected to be 100m resolution). Model outputs include uncertainty, so we will invest in automated data communication to ensure they are communicated effectively.

Data users in government agencies, local authorities, conservation NGOs and business-relevant organisations (such as Local Enterprise Partnerships, utility companies and major land owners) are a primary group of stakeholders that will benefit from our project. We will engage with representatives from these organisations through our co-design process to ensure that our outputs meet their needs. Currently, data users have information on species records at the 1-10km resolution, which is useful for large-scale strategic needs. Data users can also commission site-based surveys to meet specific operational needs. However, there are a range of needs for data that need to be met through comprehensive fine-scale biodiversity information, as will be delivered through this project. Natural capital assessments are being developed at local to regional scales (as well as nationally), and planning policies (in England, but with equivalents in other countries) need to deliver net biodiversity gain and coherent ecological networks. Our delivery of accessible, high quality information supports the vital inclusion of biodiversity in these plans.

Some of the local users (e.g. from local economic partnerships or local councils) will not be as familiar with biodiversity model outputs as users in national government agencies, so communication of results and their uncertainty needs to be effective (through visualization and text) and, with models being updated in near real time, the communication needs to be automated. We will work with data users through our co-design process to ensure that these needs are met. We will also make our outputs available via a data lab, to enable easy access to outputs, without users requiring a large data infrastructure. In our project, we have chosen to work in two target regions with project partners to focus our engagement with local data users. Our project partners already have good networks in these regions and we will link to local economic partnerships, councils, agencies and businesses who have need for biodiversity data.

A second group of stakeholders who will benefit are citizen science recorders. Many tens of thousands of people voluntarily provide biological records each year in the UK providing millions of pounds worth of contributions. Our development of Adaptive Citizen Science through Intelligent Digital Engagements will enable their volunteer resource to used more effectively and intentionally, and by providing better feedback, aligned with their motivations, will support an even more engaged citizen science base, better able to support the increasing demands for high quality biodiversity information.

Our approach of using adaptive sampling to provide bespoke recommendations to citizen science recorders in order to target their effort to the times and places that will be most informative is novel and has only been considered in a couple of other projects in the world. This will be the first large-scale test of this approach and so our increased understanding of the motivations of citizen science recorders, combined with evaluation of their response to recommendations from adaptive sampling will benefit others designing and running citizen science projects in environmental science and beyond. In our team, we represent high profile citizen science (iSpot, iRecord and the Biological Records Centre), so we are well-placed to implement the learning and tools from this project, ensuring a good legacy of of this investment.