The CLIMASOMA project aimed to contribute to an alignment of research strategies connecting agricultural management, soil structure and climate adaptation potential through its summary of the literature, its meta‐analysis and its identification of knowledge gaps. On the one hand, we focused on soil management and cropping systems, but we also investigated the current understanding of farmers’ perception of climate change, its associated risks and opportunities, related EU policy instruments and how this may influence their decisions regarding implementing soil adaptation measures. We also investigated the potential of new research tools such as natural language processing, meta-analysis and machine learning to increase our ability to extract information from existing literature and derive context-specific information from it.

Consensus on effective agricultural practices

There is considerable degree of consensus in the literature on the effects of soil and crop management practices. For three combinations of practices the scientific evidence was particularly clear: continuous living cover, organic amendments and reduced or no-till systems. Maintaining a continuous living cover on the soil in space and time is the most effective way to foster soil structure. Such a healthy soil structure is typically better at infiltrating and retaining water. Scientific research on organic amendments for agricultural land has focused extensively on the effects of biochar, even though this technique has not (yet) been widely adopted in the agricultural sector. Biochar has a positive effect on soil hydrological functioning. From the few available studies, it is also clear that the addition of organic amendments in general improves soil structure and plant available water and reduces runoff and erosion. The evidence on the effects of reduced or no-tillage shows mixed results, with adverse effects on bulk density, despite improvements in soil structure and available water. Both positive and negative effects on runoff have been reported. Although significant trends are visible in the literature, it is still sometimes difficult to differentiate among pedo-climatic regions and other context-specific factors. Many meta-analyses perform little context-specific analyses, so the information is lacking in the final publication. In addition, there is often a lack of information in the original studies used to extract the necessary meta-data. It is therefore important for studies on agro-ecosystems from all disciplines to agree on standards concerning the required meta-data reported in peer-reviewed studies.

Key policy instruments

Public policies play an important role in farmers’ decisions influencing sustainability of crop production. A set of economic and regulatory incentives are provided trough the CAP to promote more sustainable soil management and climate change adaption. The understanding and consideration by the farming community of sustainable soil management practices and climate risk management instruments in general and agricultural insurance in particular can be improved. A key opportunity to increase knowledge of the benefits of those systems is offered through rural development training, knowledge transfer and management exchanges initiatives and the inclusion in the farm advisory service, helping farmers to implement appropriate solutions for their specific situations, including aspects of climate change adaptation.

A diversity of farmer types

The more aware and concerned farmers are about the impacts and risks of climate change the more likely they will adopt an adaptation strategy. Nevertheless, farmers have both climatic and non-climatic reasons to change their practices. ‘Multi-purpose’ adaptation strategies that cover the many areas of risk and opportunities that a farmer is facing show most potential, since soil-related adaptation strategies were not always the entry point for farmers in the context of climate change adaptation. In addition, the “farmer” as such with his/her particular behavior does not exist. Grouping farmers in a particular location, based on objective characteristics of their farm in combination with their personal characteristics may help to identify potential underlying factors driving farmers climate change adaptation decision making. The communication of climate adaptation strategies needs to emphasize the connection between extreme weather events and climate change to enhance farmers’ understanding of the risks or opportunities that climate change poses. Moreover, we have to communicate to farmers the benefits soil management can bring about in the context of the weather changes they are experiencing directly. Co-learning exchanges between farmers, and with scientists, are key to increase in adopting climate adaptation practices.

Approaches to assess the direct impact of climatic drivers

Field manipulation experiments are powerful tools to understand the adaptative response of soils to climate change because they clarify the cause-and-effect relationships in both short- and long-term. The space-for-time substitution approach allows to simulate climate change scenarios ranging from current climate to altered temperature (warming) and precipitation regimes (droughts and floods).

For instance, long-term increase of precipitation affects soil hydraulic property, by reducing infiltration rate and influencing water retention, especially during dry summer. This can negatively affect microbial abundance and activity. Similarly, also below-ground C and N cycling are highly influenced by changes in precipitation regimes. Particularly, higher precipitation accelerates soil organic matter decomposition, while drought events reduce the decomposition due to the increased physical protection of soil organic matter. Bacterial and fungal communities are less influenced by decreased precipitation. Some manipulation experiments, investigating the combined effects of climatic drivers and enhanced CO2 concentration on soil properties, highlight that they soil water and microbial processes are directly linked to soil organic matter decomposition.

In general, this synthesis provides a clear overview of the current knowledge on the adaptative response of key soil properties and functioning to climate change based on field approaches. It advances our understanding on the impact of climate drivers and enhanced CO2 concentration on soil properties and functioning mainly related to grassland land use. Overall, more insights need to be obtained for other land use and soil management practices. This is of crucial importance to help us understand the complex mechanisms underlying the responses of soil system to climate change that are still not completely known.

Climatic drivers affect near-saturated hydraulic conductivity

We confirmed significant correlations between climate variables as well as the elevation above sea level with saturated and near-saturated hydraulic conductivity. While it seems very likely that these variables influence soil physical properties, the exact underlying mechanisms need to be investigated in future studies. We found indications that specific soil management practices lead to changes in saturated and near-saturated hydraulic conductivities, which were increased under perennial cultures and decreased for no-till arable fields with annual crops and for compacted soil. Our data also confirmed that it is fundamental to take the time of the measurement after the last tillage operation into account to understand relationships between soil management and saturated and near-saturated hydraulic conductivity. All management impacts turned out to be dependent on the pedo-climatic context, as they only could be observed if variations in the latter were ruled out. We found that the data availability for tension-disk infiltrometer data was too scarce and riddled with too many gaps for detailed analyses of other soil management impacts, more specific pedo-climatic context dependencies and publication bias. Furthermore, we detected indications that the available data was afflicted with experimenter bias. Altogether, it was not possible to predict saturated and near-saturated hydraulic conductivities from the available data for new sites, which echoes results of similar attempts to build respective pedotransfer functions. More measurements with better documented meta-data and better suited predictor variables would be needed for progress in this field of research. Studies quantifying soil structure evolution with respect to season, land use and soil management using X-ray imaging may turn out to provide useful insight that may point towards for more appropriate proxy variables for the saturated and near-saturated hydraulic conductivity of soil.

Natural language processing helps to extract data from vast amount of literature

At the image of this project, synthesizing a body of scientific literature over a topic and extracting structured information from it remains a task that necessitates at lot of effort and need to be frequently repeated to “keep up” with novelties. As the body of scientific is growing, so is the effect needed to synthetize evidence. While natural language processing methods cannot completely replace human interventions, they can, however, be of great help to extract specific information to build database. They can also help to identify topics among a set of documents and summarize relationships identified between a driver and a variable based on abstracts. Extracting information from PDF format or table remains an obstacle, as well as the manual tuning needed to adjust the algorithm. Overall, natural language processing methods are a great tool to support regularly updated evidence synthesis or large body of literature.