Monitoring soil carbon: The state of the art
Matt Aitkenhead is a soil scientist working at the James Hutton Institute in Aberdeen, Scotland. He works on the monitoring and mapping of soil properties, using novel digital techniques. Currently, he is focusing on finding ways to identify land management solutions that balance agricultural productivity and environmental quality.
Aidan Keith is a soil ecologist at the UK Centre for Ecology & Hydrology. As well as having a deep curiosity for soil biodiversity, he is interested in the development of methods to measure soils at scale and data synthesis to improve knowledge of land use and climate impacts on soils, with the aim of shaping more sustainable and resilient multi-functional landscapes.
Why is it necessary to monitor soil carbon? There are several reasons, both old and new. In the last 150 years, soil organic matter (which is 50-60% carbon) has gained greater acknowledgement as a vital ingredient of soil health. Along with recognition of the many functions that organic matter performs, there is an increased desire to measure how much exists throughout the soil profile.
Originally, this came from land managers who wanted to know whether their soil was ‘healthy’, so they could make decisions about inputs, ploughing and crop selection. We put ‘healthy’ in inverted commas because there are many different ways in which soil health can be defined: a soil that is good for one purpose might not be good for another. In almost all agricultural situations, however, more SOM (soil organic matter) is better than less.
More recently, the amount of SOM (and therefore carbon) in agricultural soil has become important for another reason. Globally, soils hold nearly 3000 gigatons of carbon, while the atmosphere has less than 1000 gigatons (for the moment). Soils have lost approximately 150 gigatons of carbon since human agricultural development so it makes sense that, if managed differently, they could significantly impact greenhouse gas concentrations.
Lots of research is looking into how much carbon soils could hold and how to achieve this, but there is a potential to reduce our atmospheric CO2 concentration (currently at 415 parts per million) by 50-100 ppm. Given that around 280 ppm would return us to a pre-climate change atmosphere, there’s a strong argument for trying to achieve this. This doesn’t even consider the additional improvements of increased soil carbon on our food productivity, water supply, biodiversity and other environmental benefits.
In the UK with its 5-6 gigatons of soil carbon, farmers for the last 80 years have been incentivised for agricultural productivity through various subsidy/payment schemes. These systems have resulted in a trebling of productivity, in terms of yield per unit area. One of the downsides of this intensification is that 40-50% of UK soils are at risk from erosion and compaction, both of which lead to SOM loss.
This is likely to change in coming years, influenced by public opinion, science and government legislation. Post-Brexit, the UK and Scottish Governments will be developing new agri-environment schemes, and these will likely contain mechanisms for farm payments that incentivise broader environmental benefits. If this happens, it would be impossible for farmers to achieve improved environmental conditions without considering soil carbon and SOM.
Where there are payments, there is always monitoring and verification, so any agri-environment scheme directly or indirectly targeting SOM will have to include a framework for monitoring. This will require accredited methods that can be verified and reported on. Currently, existing soil monitoring in the UK is fragmented, patchwork and inconsistent in approach. Improvements are needed, for both sensor technology and reporting systems.
We cannot rely entirely on field sampling and laboratory-based analysis as this would be prohibitively expensive, costing more than the subsidies themselves (although these are always going to be needed to provide standardisation of other methods). So new and existing technologies need to be developed and explored, to determine their potential for monitoring the amount and rate of change of carbon in UK soils.
The NERC Soil Security Programme, coordinated by Professor Chris Collins (University of Reading), hosted a workshop in September 2019 in Lancaster on “Monitoring and sensing of soil organic matter in agricultural systems of the future.” The workshop explored these needs, particularly low-cost methods that allow SOM information to be integrated with other systems. This integration would enable monitoring for agri-environment schemes and provide opportunities for decision support for farmers and policymakers.
Two of the biggest questions are (1) how much soil carbon is there across the UK, and (2) how fast is this quantity changing? Both questions need to be answered spatially, as the UK’s soil carbon is not distributed evenly – and neither is the rate of change. One of the first points agreed was that we can already answer the first question. Mapping of existing soil carbon stocks is a relatively mature subject, with several approaches that can be applied depending on scale (field, catchment or the whole country).
The priority, therefore, is to find ways of measuring how rapidly soil carbon is changing. There are a couple of issues here. Firstly, current soil monitoring programmes in the UK detect change over timescales of 10 years or more; whereas any agri-environment scheme involving payments would need change information over much shorter periods (1-3 years). Secondly, soil carbon change from one year to the next is likely to be within the margin of error for even the best existing techniques. So we need to find ways of picking up on smaller changes in soil carbon content, and doing so repeatedly from year to year.
Sensors are an important part of the solution to these issues. Multiple available technologies capture soil data at different spatial scales, costs and ease of use. These fall into four different ‘types’: (i) remote sensing (e.g. satellite imagery); (ii) spectroscopy (think the tricorder from Star Trek); (iii) smartphone-based apps for soil profile image analysis and (iv) probes that measure properties directly (e.g. temperature, moisture, pH).
Much of the challenge for sensor technology is not to do with the sensors themselves, but how to handle the data they produce. For satellite and ground-based imagery, the quantity of data available means that processing and interpretation has become the bottleneck to monitoring. Another issue is the shortage of up to date field data, without which all this imagery is just pretty pictures with no meaning.
Integration of data from multiple sensors is another issue. Most research in soil sensing up to now has involved development of stand-alone methods for tackling specific questions. What is needed are ways of bringing data from these methods together to improve on how these questions are answered.
Sometimes this can be as simple as integrating different formats or data from different locations or dates. Often however, the problem is that researchers do not know what data is already available or are not able to share data effectively. A centralised soil data ‘clearing facility’ would help greatly with this, but even better would be encouragement for researchers to upload their data to such a facility. The NERC EIDC (Environmental Information Data Centre) is intended for such a purpose but is of limited use if people are reluctant to share or find it difficult to do so.
In part, the problem of data sharing is due to not having the time set aside within projects for the proper formatting, standardisation and organisation of generated data. Research funding is still prioritised towards the generation of new data and results, not towards making sure that this data is useful for future, as-yet-unspecified purposes.
It turns out that sensor technologies are probably the least of our worries – standardisation, integration and governance of a coherent system that everyone can use are the real goals. To accomplish these, a coordinated and centralised investment is needed. Standardised data collection methods and instrument calibrations would be the first big win from this, leading to improved data sharing and re-use.
Data governance and licencing could also be tackled within this system, giving data ‘providers’ an assurance that they would get credit for their work and giving data ‘users’ more confidence that the data meets a specified standard.
Once researchers have SOM data, how could they use it better? A first, important step would be to understand the impacts of specific management practices, under different conditions, on SOM. To do so requires modelling and statistical analysis of the data, which allows us to say what will happen in places where monitoring has not been carried out. It also improves our understanding of why certain things happen, which is much more useful than simply knowing that they happen.
Once models have been successfully developed, we can produce maps of what will happen where, when it will happen and how. It can also answer the ‘which’ and ‘who’ – who needs to act to achieve a desired outcome, and which action they should take. Development of novel modelling and analysis approaches are therefore just as important as the data itself.
If we can start answering these questions, scientists can start informing policy in a way that leads to greatly improved land management and environmental conditions in the UK. Fundamental to this is the translation of knowledge about soil carbon into language that farmers and policymakers understand. Without this, meaningful change is impossible.
Achieving this translation requires that soil scientists step out of their comfort zone and involve social scientists directly in their work – not once the ‘science’ has been accomplished, but at the outset. There are fundamental social and economic research questions about how to increase soil carbon within the policy, economic and regulatory landscape of UK agriculture. Without tackling these questions, the biophysical science will be ineffective.