Work Packages

WP1: Project Administration 

 I•GIS, AU

This WP will cover the communication with IFD, oversee the progress of each WP and coordinate the project as whole (I-GIS), arrange internal project seminars among the partners (SkyTEM, RM), and manage accounting (AU AGRO). Internal project administration in the partner organization is covered for some partners in WP1 budget due to different approaches in budget setup at timepoint of application. WP1 only covers internal administration and dissemination, not external.  

WP2: Geophysical mapping methodologies

 SkyTEM

Geophysical mapping instruments According to Minasny et al. (2019), gamma radiometrics and electrical conductivity data constitute the best predictors for delineating peat extent. Therefore, the accuracy of the C mass estimate will be significantly improved by combining gamma radiometrics data and TEM data which can characterize the geological/soil settings and groundwater table. The geophysical equipment will be developed to be sling-load or operated under a suitable unmanned aerial vehicle (UAV). The UAV will allow surveying with gamma ray instrumentation, GPR and TDEM equipment making an easy switch between the three during field work. Automation of the workflow of data from field to cloud including automatic quality control (QC) will benefit the operational aspects and economy. A major development will be to model, design and develop a new innovative high bandwidth, 100 Mhz TEM (SkyTEMPico) transmitter and receiver platform that will push the envelope of TEM method to map the upper few meters of peat and soil. It will also improve the resolution and ability to use on-time measurements. With a smaller platform, the ability to operate at 1-3 m above ground level and higher repetition rate will improve the lateral resolution compared to other known systems.  

WP3: Data Acquisition

 AU-Agro, SkyTEM, Region Midtjylland

Data acquisition Environmental data (covariates) and soil observations are required as input data to model the target variables (peatland type and extent, peat thickness and groundwater depth) in detail. Typical environmental data consist of legacy data on landscape, soil maps, historical land use maps, drainage systems related data, and remote sensing data, such as satellite images and digital elevation models. Early in the project, the wetlands and river valleys will be classified and subdivided according to valley type and lithology. This classification will then help evaluating the spatial variability within each type and defining how many soil samples should be collected by type. New environmental data will be collected both with proximal and remote geophysical sensors for the study areas. For model development and validation, soil observations will be collected in the field (e.g. peat thickness, field pH, conductivity at different depths) and samples will be analyzed in the laboratory (e.g. bulk density, C content). Based on the preliminary analysis of the collated data, a field sampling design will be defined for each study area (e.g. from a conditioned Latin hypercube sampling approach; Minasny and McBratney, 2006).  

WP4: Data Integration and ML 

AAU, AU, I•GIS

Data integration and machine learning Following the DSM approach (McBratney et al., 2003), an advanced ML technique, CNNs, will be assessed to predict the target variables. CNNs being a powerful method, we expect they will yield a much higher prediction accuracy than classical ML techniques. As CNNs require the use of images as input, the gathered data will be integrated to allow image extraction, building on ongoing work (Beucher et al., 2019a and b). The project will not only produce predictive maps for the target variables, but also uncertainty maps and model interpretation, which both constitute significant gaps within current soil mapping studies. Since the predictive maps and their associated uncertainties will be used by decision-makers, it is crucial to get some insight into the developed CNN models drawing on recent work on explainability (Samek et al., 2019). 

WP5: 3D modelling

I•GIS, AU AGRO, AAU   

Cloud computing, data management and 3D modelling The methodologies and algorithms developed in WP4 will be implemented in the already existing GeoCloud (a cloud-based data management system provided by I•GIS) together with the necessary infrastructure needed for these data types to be easily read in, visualized and integrated with the 3D geological modelling software, GeoScene3D (also developed by I•GIS). The predictive models created within WP4 would greatly benefit from an external validation through 3D modelling. The predicted variables constitute physical features and visualizing all data together with the predicted features in a 3D environment will add high value in quality-controlling the predictions and in turn the ML models. In addition to allowing visual inspection of models and data combined, we will develop tools in GeoScene3D that allow experts to interact with the ML models through an API and through cloud services. We believe that combining the state-of-the-art ML models and allowing unbiased predictions on large datasets with the insight, expertise and experience of a trained specialist will be key to achieving the best possible mapping of these peatlands.  

WP6: Effects and scenarios

I•GIS, RM, AU AGRO  

Scenario estimations Having a 3D model of peat, together with the groundwater depth and information on the relation between CO2 emissions and drained peat, we will build a concept to evaluate measures that could be applied to a specific area. This concept will be able to compute the potential net reduction in CO2 emissions based on different proposed scenarios, such as no action, stopping agricultural actions or flooding parts of the land. The exact scenarios to be included and the development of the methods to compute the results of the different scenarios will be established in the project. A web-based frontend will be developed as well as an API to deliver the results from this project to be utilized in decision-making tools for multifunctional land consolidation. 

WP7: User-involvement, capacity building and outreach

RM, AU AGRO, (I-GIS, SkyTEM) 

To secure that the project outcomes reach key stakeholders, local authorities, the scientific community, and are implemented in the society as a whole, this work package will focus on creating focus groups, arranging workshops and seminars, and will disseminate results through national and international conferences and internationally recognized journals.