Analytics for Digital Earth - Workshop 2
Date: 27th May 2022
Location: ARC / Online
Speakers: The subject for our first hybrid workshop was soil carbon with around 40 people attending in person and online. The first speaker was Prof John Crawford (University of Glasgow) who spoke about global soil health and how we can improve farming practises through digital technology, policy, and regulation. Next, we had Prof Pete Smith (University of Aberdeen) talk about how to measure, report, and verify soil carbon change. Finally, Dr Rachel Opitz (University of Glasgow) talking about understanding human-environment interactions with soils over different time scales from an archaeological perspective.
Discussion Format: There were two breakout discussion sessions, with the group split into an in-person groups and a couple of online groups for each session. The first session followed on from Prof Crawford’s talk on soil health and the questions he posed. The second session looked at the questions posed by Prof Smith and Dr Opitz’s talks with an additional focus of looking at how digital twins could be implemented in this context.
Session 1: The discussion in session 1 focused on global soil health and how we deal with the uncertainty, with topics including:
- How to predict value propositions and confidence?
- How to avoid unintended consequences?
- How do we provide confidence on what we try to predict?
- How do we measure impact at scale in a cost-effective manner?
- How do we accelerate scientific discovery and translation to impact on the ground?
Session 2: The discussion in session 2 was initially focused around measurement, reporting and validation of soil sampling, with specific questions about:
- How to develop sampling schemes which scale at field, to farm, to landscape and beyond?
- How to account for the different characteristics of the measurement processes?
- Can we develop a citizen science approach to collecting soil samples?
- How can we combine different measurements collected using different methods?
The session had a particular interest in experimental design in the context of placement of soil sensors. Specific questions looked at:
- Can we effectively design where future measurements are taken based on existing measurements?
- Can we quantify how effective our soil sampling schemes are?
- Can we develop a model to understand soil carbon which can also account for archaeological impacts on the soils?
Discussion Focus points:
- Collecting Data
- Finding a solution that means that those who provide data benefit from the process and therefore collect a larger amount of reliable data
- Quantifying which sources of data are reliable and therefore how cost effective they are
- Data sources are fragmented, we need a systematic approach to bringing them together
- Citizen Science
- Citizen science could be a hugely valuable source of data, but it is difficult in terms of reliability and collecting data in the most effective places- sampling schemes may be unrepresentative
- Design of Experiments
- Collecting sample and placing sensors needs to be done in an intelligent way
- Designing effective methods for a variety of scenarios is important
- Can we use less reliable sources of data to helps us identify the location for more precise (expensive?) sampling methods
- Combing Multiple Data Sources
- One of the biggest challenges of digital twins appears to be integrating and combining multiple sources of data and/or models
- Aggregating different data sources seems possible, but understanding uncertainty in these contexts seems more difficult
- Understanding Co-Benefits
- Need to understand how to quantify co-benefits
- For soil carbon this includes: water storage, reduction in flood and drought risk, biodiversity impacts, social impacts, at the same time as avoiding unintended consequences.
- Bringing People Together
- Creating relationships before beginning projects is potentially important, can help with cooperation and also manage expectations
- This can also help take advantage of co-benefit scenarios
The way forward for the next workshops: Our next workshop will focus on water with areas of interest involving monitoring rivers and lakes (water quality), water quantity (through sensors and earth observation), specifically flows and extremes (floods and droughts) and water infrastructure (wastewater treatment, CSO discharges). Future workshops will take the learning from the preceding workshops and begin to tackle some of the common/shared analytical challenges as we move to considering how to construct a digital twin.