Environment, Natural Resources and Agriculture research strategy 2027-2032: consultation analysis

11. Question 9

Theme: Impact Routes – Decision Support, Living Labs and innovation

Question 9: What barriers exist to delivering effective data analysis and modelling in the current ENRA Research Programme?

Introduction

The majority (42, 60%) of consultation respondents answered question 9.

The vast majority of respondents identfied barriers to effective data analysis and modelling in the current ENRA research programme

Theme 1: Data accessibility

The majority of respondents (all respondent types, in particular Research Institutes and Centres of Expertise and other public bodies) identified access to data as an existing barrier to delivering effective data analysis and modelling in the current ENRA research programme. The availability of datasets, analysis and modelling to inform policy decision was commonly described as ‘fragmented’ and ‘siloed.’

It was further noted that existing datatsets can vary significantly in terms of accessibility and quality. This led to calls from a small number of Research Institutes and Centres of Expertise for a ‘central repository’ for data and associated software. However, it was also noted that curation and synthesis of relevant datasets can take significant time and resources which can limit the extent to which research outputs can inform time-senstive policy decisions. Respondents also noted that access to data can be further complicated by General Data Protection Regulation (GDPR), particularly for smaller and potentially disclosive datasets.

Some respondents noted that the inaccessibility of data can be exacerbated by a reluctance to share datasets, analysis and modelling, particularly among SMEs and community groups. To this end, these respondents welcomed the focus on fostering a culture of collaboration and active engagement which could support greater access to relevant and robust datasets, analysis and modelling for different stakeholders.

While a few organisations noted that there has been progress with the adoption of FAIR data principles, feedback suggested this was uneven and that a more strategic approach was needed to ensure widespread adoption in cross-sector analysis and modelling.

“Data archiving and accessibility is also likely to be a barrier to this. There are some excellent examples of how data has been made available to the wider community through FAIR principles and initiatives to make data accessible, such as EoRNA^[3] and EoRNA2, but too many datasets are not made available in a FAIR way making them unavailable for decision support development and modelling. One of the main barriers to delivering effective data analysis and modelling in the current ENRA programme is therefore fragmentation. Projects and themes within the Strategic Research Programme often operate in isolation, which makes it difficult to build integrated decision-support tools that draw on the full range of available evidence. While there are good examples of progress, these remain isolated rather than systemic.” The International Barley Hub

“A central repository for data and associated software would greatly aid access to data collected. This data needs to be provided in an agreed format and ontology for describing the data, alongside a maintained list of datasets, tools etc., for re-use. There is also a lack of high-quality data for some food system sectors, particularly the bakery sector and other more extensively processed foods. This would require work with stakeholders to collect this information but would really aid future modelling of the food system and food supply chains.” Rowett Institute

“There would appear to be ‘closed minds’ among some researchers as to which models are the most appropriate for a particular challenge. An initial part of the programme needs to break down any barriers which exist between modelling teams.” Hannah Dairy Research Foundation

Theme 2: Lack of data governance mechanisms

Many respondents (all respondent types) reported that a lack of data governance in the current SRP was a barrier to delivering effective data analysis and modelling. These respondents said a lack of data governance risks a lack of shared understanding, limited data integration and quality assurance which could lead to less impactful and a more complex data landscape.

They suggested that agreed formats and ontologies would be essential to ensure greater standardisation of datasets, facilitate greater collaboration and avoid duplication of effort. Although, it was also noted that data governance should not undermine current approaches to accessing data.

“Barriers include fragmented data systems, inconsistent data standards, and limited capacity for cross-sectoral modelling. Addressing these through improved governance and collaboration will enhance the effectiveness of data analysis.” Seafood Scotland

“There are currently gaps in relation to the standardised metrics that are needed to model all of the outcomes for health, sustainability and resilience in our food system that the ENRA research programme is seeking to address. These issues underline the need for this strategy to play a more prominent role in underpinning capacity and the co-ordination of data generation in this space; with clearer read across to existing policy outcomes such as the Good Food Nation Plan and the Scottish Dietary Goals.” Food Standards Scotland

Theme 3: Skills, capacity and resource constraints

Many respondents (all respondent types) highlighted skills, capacity and resource constraints as existing barriers to delivering effective data analysis and modelling in the current SRP. Respondents said these issues could be more common for rural stakeholders, smaller NGOs and community groups who often lacked the required tools, training and technical capability to engage and contribute to data-intensive research (in terms of scale and complexity).

Some capacity issues were also reported for the Research Institutes/Centres of Expertise (by all respondent types) where gaps in advanced analytical and modelling expertise (excluding Biomathematics and Statistics Scotland) has limited the further development of robust and policy-focused tools. It was suggested that this has impacted on certain areas, such as soil monitoring, where high costs and a small pool of experts has restricted the volume of data that can be collected and analysed.

Some respondents noted that capacity was further constrained by tight delivery timescales with Research Institutes and Centres of Expertise reporting difficulty in engaging with academics in their own and other institutions, particularly those early in their careers and when projects require significant input over a short period of time.

In some responses, capacity issues and constraints led to calls for sustained investment in training and ongoing support to ensure greater adoption of advanced data modelling.

“Institute capacity for delivering modelling tools is most often at technology readiness level (TRL) 6 (System/subsystem model or prototype demonstration in a relevant environment). TRLs higher than this may mean that alternative processes and/or funding may be needed. The TRL need may also be influenced by client capacity such that greater efficiency and impact is achieved by using models with clients rather than providing model to clients.” James Hutton Institute

“Funding for soil monitoring in Scotland is limited. Collecting data to assess soil health is expensive and there are limited numbers of experts to analyse this data.” SEPA

“It is a particular barrier for CXC and the CoEs to deliver effective data analysis and modelling that we operate on very short timelines. Academics often find that they are unable to engage on our timelines due to prior commitments. This is in particular an issue where we are looking for a considerable time commitment over a short timeframe from early and mid-career researchers.” ClimateXChange

A small number of respondents did not identify any barriers

A small number of respondents (two, 4%) did not identify any existing barriers to effective data analysis and modelling. One respondent stated that there were no existing barriers to effective data analysis and modelling while the other respondent was unsure.