Laboratory for agricultural decision support systems (LADS)

There is currently no open algorithm for optimising N rates to grain crops for general and free use. The aim of this the proposed project is to develop public algorithms between multispectral reflectance measurements of the crop and the economically optimal nitrogen rate.

We aim to facilitate optimisation of nitrogen fertilisation for most of Sweden's grain producers. To achieve this, we will develop and evaluate prediction models for economically optimal nitrogenrate (EONR) in wheat (Triticum aestivum L.) and malting barley (Hordeum vulgare L.), based on near real-time satellite reflectance data (Sentinel-2) at the time of supplemental fertilisation.

The prediction models will be designed for use in satellite-based decision support systems for precision agriculture (e.g. CropSAT), to translate satellite data into directly applicable N rate recommendation maps. Better-optimised nitrogenrates mean better profits for farmers, better scope for achieving quality goals for produce and reduced environmental risks in terms of nitrogen leaching and/or volatilisation of nitrogen compounds.

Project group

• Kristin Piikki, SLU
• Mats Söderström, SLU
• Henrik Stadig, Hushållningssällskapet Skaraborg

Project period

The project starts in 2019 and ends in 2021.

Funding

Stiftelsen lantbruksforskning (contract: O-18-20-162).

Publications

Web pages, applications

• Automatisk kvävefördelning i höstvete i CropSAT med Target-N-algoritmen (in Swedish)

Well-informed decisions on land use and agricultural practices are crucial for food security and sustainable agriculture. Agronomic decisions must be based on local conditions and this call for the development of efficient soil mapping methods of East African agricultural soils.

The present project deals with pedometric considerations and outline methodologies for in situ and ex situ soil measurements by use of proximal sensor, local map adaptation of large-scale digital soil maps and inclusion of soil data in crop suitability modelling. The work focusses on soil properties that are important for crop productivity (the content of organic carbon, texture, plant-available nutrient content, and soil pH) and results are intended for use by agricultural extension officers, technicians and other scientists. An important part of the project is to disseminate the results to stakeholders and educate agricultural extension officers on how to use the digital soil map information in management decisions in smallholder farming systems in East Africa.

The project is a collaboration project between the International Centre for tropical Agriculture (CIAT) and SLU.

Various sensors (a-e) used in field work in Embu, Kenya

Project leader

Kristin Piikki, SLU/ CIAT

Project period

The project started in 2014 and ends in 2019.

Funding

Formas/SIDA (contract: 2013-01975)

Publications

Peer-reviewed journal articles

• Nijbroek, R., Piikki, K., Söderström, M., Kempen, B., Turner, K., Hengari, S., & Mutua, J. (2018). Soil Organic Carbon Baselines for Land Degradation Neutrality: Map Accuracy and Cost Tradeoffs with Respect to Complexity in Otjozondjupa, Namibia. Sustainability, 10(5), 1610.
• Piikki, K., Winowiecki, L., Vågen, T. G., Ramirez-Villegas, J., & Söderström, M. (2017). Improvement of spatial modelling of crop suitability using a new digital soil map of Tanzania. S. Afr. J. Plant Soil, 34, 243-254.
• Söderström, M., Piikki, K., & Cordingley, J. (2017). Improved usefulness of continental soil databases for agricultural management through local adaptation. South African Journal of Plant and Soil, 34(1), 35-45.

• Piikki, K., Söderström, M., Eriksson, J., Muturi John, J., Ireri Muthee, P., Wetterlind, J., & Lund, E. (2016). Performance evaluation of proximal sensors for soil assessment in smallholder farms in Embu County, Kenya. Sensors, 16(11), 1950.

Conference papers

• Can global soil organic carbon maps be used in policy decisions on practical agricultural management? Global symposium on soil organic carbon. FAO Rome, Italy, 21-23 March 2017
• Development of a soil organic carbon baseline for Otjozondjupa, Namibia. Global symposium on soil organic carbon. FAO Rome, Italy, 21-23 March 2017
• The Importance of Soil Fertility Constraints in Modeling Crop Suitability Under Progressive Climate Change in Tanzania. Agriculture and Climate Change - Adapting Crops to Increased Uncertainty (AGRI 2015), Amsterdam, The Netherlands 15-17 February, 2015

Technical report

• Soil carbon under current and improved land management in Kenya, Ethiopia and India: Dynamics and sequestration potentials. Working Paper. CIAT Publication No. 475, Nairobi, Kenya. 2018.

Blog posts newsletters etc.

• Pedometri i praktiken: utveckling av beslutsstöd för hållbar intensifiering i småskaliga odlingssystem i Östafrika. SLUs kunskapsbank, 2018
• Sequestering carbon in African soils – we need data, knowledge and tools for decision support. NEWSLETTER#2 2018 - “4 per 1000“ Initiative, page 7.
• Odling med större precision på Mt. Kenyas sluttningar. Resurs, 2016
• Odlingsförutsättningar i Östafrikanska jordar kartläggs med satellit. SLU, 2013

Map books / Web map applications

• Shikomoli village, Western Kenya, 2018: Map Book of Shikomoli / Web application
• Mukuyu village, Western Kenya, 2018:  Map Book of Mukuyu / Web application

R packages

• MapsRinteractive
• SurfaceTortoise

Datasets

• Pilot Project Land Degradation Neutrality (LDN), Namibia: Establishment of a baseline for land degradation in the region of Otjozondjupa, CIAT, 2017

Video

• Very short video clip by Mats Söderström, SLU from two farmer’s workshops in western Kenya, 9 and 11 Jan  2019

– integrating data from satellites and N-sensors for nitrogen recommendations and crop status mapping within fields

The long title is the name of the initial SLF-funded project that lead to the development of the web application CropSAT. CropSAT has become a widely used decision support tool primarily for precision application of nitrogen. In 2018 the system had more than 20000 users, most in Scandinavia.

Responsive fertilisation of winter wheat (Triticum aestivum L.) is often adopted, with nitrogen (N) applied two or three times between the developmental stages of tillering and booting. Satellite-based decision support systems (DSS) providing vegetation index maps calculated from satellite data are available to aid farmers in adjusting the topdressing N rate site-specifically to the current season and to variations in growth conditions within the field. In this project, the freely available CropSAT DSS was developed and evaluated. The system provides farmers with raster maps of the modified soil-adjusted vegetation index (MSAVI2) or normalized difference vegetation index (NDVI) (the latter in Denmark, the former elsewhere) calculated from data obtained from satellites Sentinel-2 (ESA,EU).

The project was a collaboration project between SLU, Hushållningssällskapet, DataVäxt AB, Agroväst Livsmedel, and Greppa Näringen (Focus-on-Nutrients). Later on SEGES in Denmark became involved in the development.

Further research and development on functionality of CropSAT and satellite-based systems is ongoing in LADS.

CropSAT – a satellite images based decision support system, free to use. The first application of its kind in Scandinavia.

Project team

Mats Söderström, Kristin Piikki, Lena Engström, Sandra Wolters (SLU); Thomas Börjesson (Agroväst); Henrik Stadig (Hushållningssällskapet); Johan Martinsson (Dataväxt AB)

Project period

The initial project started in 2013 and ended in 2014. Later external support from other donors and continued in project form until 2018, when it was acquired by DataVäxt AB. Research and development on CropSAT and satellite-based systems continuous in LADS.

Funding

Swedish Foundation for Agricultural Research (SLF) (proj. no. H1233115), Västra Götalandsregionen och SLU (projekt: RUN 2018-00141) and Focus-on-Nutrients (Swedish Board of Agriculture). Development have been supported by SEGES (in Denmark) and Yara (in Norway).

Publications

Peer-reviewed journal articles

• Wolters, S., Söderström, M., Piikki, K., Reese, H., Stenberg, M. 2021. Upscaling proximal sensor N-uptake predictions in winter wheat (Triticum aestivum L.) with Sentinel-2 satellite data for use in a decision support system. Precision Agriculture, published online
• Söderström, M., Piikki, K., Stenberg, M, Stadig, H., Martinsson, J. 2017. Predicting nitrogen uptake in winter wheat by combining proximal crop measurements with Sentinel-2 and DMC satellite images in a decision support system for farmers. ACTA Agric. Scand. Sect. B, Soil and Plant Sci., 67, 637–650.

Conference papers

• Alshihabi O., Piikki K., Söderström M. 2020. CropSAT – A Decision Support System for Practical Use of Satellite Images in Precision Agriculture. In: El Moussati A., Kpalma K., Ghaouth Belkasmi M., Saber M., Guégan S. (eds) Advances in Smart Technologies Applications and Case Studies. SmartICT 2019. Lecture Notes in Electrical Engineering, vol 684. Springer, Cham.
• Börjesson, T., Wolters, S., Söderström, M. 2019. Satellite-based modelling of protein content in winter wheat and malting barley. In: Stafford, J.V. (Ed.) Precision Agriculture '19. Wageningen Academic Publishers, Wageningen, the Netherlands, pp 581-588.
• Wolters, S., Söderström, M., Piikki, K., Stenberg, M. 2019. Near-real time winter wheat N uptake from a combination of proximal and remote optical measurements: how to refine Sentinel-2 satellite images for use in a precision agriculture decision support system. In: Stafford, J.V. (Ed.) Precision Agriculture '19. Wageningen Academic Publishers, Wageningen, the Netherlands, pp 1001-1008.
• Söderström, M., Stadig, H., Martinsson, J., Piikki, K., Stenberg, M. 2016. CropSAT – A public satellite-based decision support system for variable-rate nitrogen fertilization in Scandinavia. 13th International Conference on Precision Agriculture (ICPA), St Louis, MI, USA

Technical report

• Söderström, M., Stadig, H., Nissen, K., Piikki, K. 2015. CropSAT: kväverekommendationer och grödstatuskartering inom fält genom en kombination av satellitdata och N-sensorer. POS Teknisk rapport 36, Institutionen för mark och miljö, Sveriges lantbruksuniversitet. (In Swedish).

Popular

• Söderström, M. 2018. Satellitbilder året om med CropSAT Arvensis, 7, 2018: 26-28. (In Swedish)
• Söderström, M. 2017. CropSAT ännu bättre i år. Arvensis, 3, 2017: 15. (In Swedish)
• Söderström, M. 2016. Hur funkar CropSAT? Arvensis, 3, 2016: 22-24. (In Swedish)

Web pages, applications

• CropSAT – gödsla rätt med satellitdata. Precisionsskolan, Verktygsportalen (in Swedish).
• Swedish version: CropSAT.se / Global version: CropSAT.com
• Automatisk kvävefördelning i höstvete i CropSAT med Target-N-algoritmen (in Swedish)

Video, pod, blog with info on CropSAT

• Kunskapspodden. 2018. Precisionsodling del 1 – från markkartering till satellitbilder. Pod (in Swedish) partly about CropSAT.
• Aronsson, H., Hoffmann, M. 2018. Precisionsodling, Goodla, Video (in Swedish) on precision application on variable-rate application of nitrogen – one part is about CropSAT.
• SEGES. 2018. CropSAT til gradueret gødskning (video in Danish).
• Agroväst Livsmedel. 2017. Agroväst 25 år – Agroväst samordnar utvecklingen av Västsvenskt lantbruk. Video (in Swedish) partly about CropSAT.
• SEGES. 2016. Kom godt i gang med CropSAT, og brug det til omfordeling af kvælstof (webinar 7. april 2016) (video in Danish).
• Sinergise, 2016. Sentinel Hub powering CropSAT.dk. Blog.

Solvi.nu is web application for management of drone images. One special feature in the system is management of images acquired over field crop trials. The aim of this tool was to facilitate the use drones in crop trials, which is of world-wide interest.     

Much of what we know about how new crop varieties react under field conditions, as well as how to optimize fertilisation and other management actions, emanates from field trials. A field trial may consist of hundreds of small plots in which crop response to management can be assessed in a statistically rigorous manner. During later years the use of crop sensors have become more common in this work, and drones (unmanned aerial vehicles, UAV) as a platform for measurement is regarded as very promising.

In order to automate the management of drone images from field trials, we have developed a new cloud-based tool that requires a very limited input from the user. The tool automatically detects boundaries of each plot in a mosaic image through a neural network model, and calculates summary statistics for individual bands and selected vegetation indices. Plot-wise data can then be exported as text files or in geographical formats for further analysis in geographical information systems.

The Zonal statistics and Trial plot extraction tool was implemented in Solvi (www.solvi.nu) - a web application for drone imagery analysis. Further development is ongoing.

Through the Zonal statistics and Trial plot extraction tool in Solvi.nu it becomes much easier to use drones in field trials.

Project team

Mats Söderström, SLU; Sofia Delin, SLU; Igor Tihonov, Solvi.

Project period

The project started in 2018, and a functioning version is currently online. Further development is ongoing.

Funding

Vinnova (dnr: 2016–04248) within the Testbed UAV project, coordinated by RISE Research Institute of Sweden AB.

Publications

Web pages, applications

• Solvi.nu – free trial is available.
• Free trial plot data showcase on Solvi from SLU

Blogs

• Solvi. 2018. Introducing Zonal Statistics and Trial Plot extraction tools. Blog.

DSMS is an open geodatabase with information on the topsoil of the arable land in Sweden. The aim is that the data shall be accurate and detailed enough to be useful for precision agriculture, i.e. for management decision within fields.

The spatial resolution of DSMS map layers is 50 m × 50 m. Currently, it covers > 90% of the arable land of Sweden (~2.5 million ha). Non-agriculture land and areas with organic soil are not covered. Access to a number of national datasets made it possible to build the DSMS. Results from two soil sampling campaigns (~15 000 samples) were combined with remote sensing data (gamma radiation data from airborne radiometric scanning and a digital elevation model derived from airborne laser scanning) plus a Quaternary soil deposit map. So far, multivariate adaptive regression splines models were parameterized for content of clay, sand and organic matter, as well as buffering capacity and target-pH. DSMS data is provided for free by the Geological Survey of Sweden (SGU). Examples of applications developed based on the DSMS are web applications used by farmers to generate prescription files for variable-rate seeding and variable-rate liming – e.g. Markdata.se.

Later research has focused on developing interactive functionality between digital soil mapping databases and the user’s local data. This means that a user can combine local data with e.g. DSMS maps, and automatically downscale the database to generate improved local maps.

Currently, research is ongoing for developing new map layers of DSMS.

DSMS is free to use and can be used for the development of web applications suitable for precision agriculture. The image shows a map for variable-rate liming developed in Markdata.se, an application based on DSMS data.

Project team

DSMS development project team: Mats Söderström and Kristin Piikki, SLU; Gustav Sohlenius and Lars Rodhe, SGU. Development of Markdata.se was carried out by Kristin Piikki and Mats Söderström, SLU; Henrik Stadig, Hushållningssällskapet; Johan Martinsson, DataVäxt AB. Current research is carried out by PhD student Karl Adler, SLU.

Projects and Funding

A pre-study was carried out by SLU in collaboration with Hushållningssällskapet in 2010-2011 (Swedish National Space Board, proj no. 199/10). The initial development of DSMS started in 2013 was also financed by the Swedish National Space Board (proj no. 214/13) and SGU. Further developments have been funded by the Swedish Foundation for Agricultural Research (SLF; O-15-20-566). Current projects for the development of DSMS are funded by Västra Götalands Regionen and SLU (VGR/SLU project RUN 2018-00141).

Acknowledgements

DSMS could be created thanks to availability of data från Lantmäteriet, SGU, the Swedish Board of Agriculture (Jordbruksverket) and Swedish Environmental Protection Agency (Naturvårdsverket).

Publications

Peer-reviewed journal articles

• Piikki, K., Söderström, M. 2017. Digital soil mapping of arable land in Sweden – Validation of performance at multiple scales. Geoderma, in press. DOI: 10.1016/j.geoderma.2017.10.049.
• Söderström, M., Sohlenius, G., Rodhe, L., Piikki, K. 2016. Adaptation of regional digital soil mapping for precision agriculture. Precision Agriculture, 17, 588-607.

Conference papers, book chapters

• Piikki, K., Söderström, M., Stadig, H. 2017. Local adaptation of a national digital soil map for use in precision agriculture. Advances in Animal Biosciences: Precision Agriculture (ECPA) 2017, 8:2, 430-432.
• Piikki, K., Söderström, M. 2016. The new digital soil map of Sweden – for free use in precision agriculture: 13th International Conference on Precision Agriculture, St Louis, MI, USA.

Technical reports

• Söderström, M., Piikki, P. 2016. Digitala åkermarkskartan – detaljerad kartering av textur i åkermarkens matjord. POS Teknisk rapport 37, Institutionen för mark och miljö, Sveriges lantbruksuniversitet. (In Swedish).
• Söderström, M. 2013. Digital markkartering av Skånes åkermark. POS Teknisk rapport 26, Institutionen för mark och miljö, Sveriges lantbruksuniversitet. (In Swedish).

Web pages, applications

• DSMS download page at SGU
• Digitala åkermarkskartan. Precisionsskolan, Verktygsportalen (in Swedish).
• Markdata.se web application for precision ag run by DataVäxt AB (in Swedish)
• FAO magnifier
• Clay content of Swedish arable land. Viewer suitable for smartphones.

Baby Grain Passport – Traceable and sustainable production of cereals for baby food | Externwebben (slu.se)