Increasing efficiency in rangeland-based livestock value chains through machine learning and digital technologies
Start: Apr 2024 - End: Apr 2025

Project summary
Pastoral livestock production on rangelands is an important land-use system and contributes between 15 and 60 % to the agricultural GDP of countries in eastern and southern Africa. Largely mobile herds exploit the temporal and spatial heterogeneity in resource availability (pulses) on rangelands. This production strategy has advantageously low fossil fuel input needs but is very knowledge and information intensive. Therefore, opportunities derived from digitalisation will have a high potential to increase efficiency (‘precision pastoralism’).
To successfully introduce the technology InfoRange uses a transdisciplinary approach to co-design the ICT solutions with users and embeds them in social innovations. By an actor- and activity oriented approach we build on the knowledge of different involved actor groups to understand how their decision-making can be improved through ICT.
Co-designed ITC solutions will enhance sustainable rangeland use and efficiency in livestock production through improved grazing management and veterinary service provision. InfoRange will combine user-generated information (e.g. similar to geotagging photos in google maps or live traffic updates) with remotely sensed data. State-of-the-art machine learning models will be developed to analyse the generated crowd data (e.g. time series), capture and understand phenomena such as differences in pasture use intensity as well as classify and recognise patterns in different scenarios. Including representatives of different governance bodies from the onset of the project permits to creation of outputs in formats suitable to enhance policy decisions.
Consortium
- German Institute for Tropical and Subtropical Agriculture – DITSL
- Department of Agricultural and Biosystems Engineering of the University of Kassel (UK)
- Center for Research and Development in Drylands - CRDD Center for Research and Development in Drylands – CRDD
- University of Nairobi (UoN)
- Namibia Nature Foundation (NNF)
- Namibia University of Science and Technology (NUST)
Objectives
InfoRange aims at improving rangeland use and governance and increasing resource-use and production efficiency in rangeland-based livestock production through digital and ICT applications/services that permit user-generated information acquisition and transmission. It also seeks to contribute to integrating external telemetry and observatory data with land-user generated data on bio-geo-physical ecosystem features in order to render digital and ICT services more relevant for land-users’ immediate management decisions on grazing, watering and health management.
InfoRange uses a transdisciplinary approach to
- Adapt, modify and further develop existing ICT tools for decision support in rangeland management and use as well as for veterinary service provision
- Develop procedures and solutions to enhance the use of ICT tools by land-users and increase their distribution and accessibility under reduced network coverage
- Combine approaches from citizen-science, crowd-data sourcing, machine-learning and participatory monitoring and evaluation to render these tools more relevant for decision-making at different governance levels
Project location
In Kenya, the project will be implemented in the northern arid and semi-arid Marsabit County in Laisamis and Moyale Subcounties – i.e. two study locations. In Namibia, InfoRange will be implemented in the Kavango East Region (including the Ndiyona Constituency) and Omaheke Region (including the Otjombinde Constituency). The chosen sites represent communal livestock-management systems, which will ensure that the techniques developed for better resource monitoring and evaluation (M&E) are relevant for the people on the ground and will enable an easy scaling-up of the project activities.
Project structure, work packages and their tasks
InfoRange is structured in seven Work Packages (WPs) with defined Work Tasks (WTs) for each WP.