It is our pleasure to announce that our project “Namibia, a new dataset for rangeland and pasture management” has been selected for funding by Lacuna Fund. Lacuna Fund is an American Institution that supports the creation, expansion, and maintenance of equitably labeled datasets that enable the robust application of machine learning (ML) tools of high social value in low and middle income contexts globally. The call, named “Labeled Agricultural Datasets for Machine Learning Solutions in Sub-Saharan Africa” was launched in 2021 and aims at “providing funds to create, expand, and/or maintain datasets that fill gaps and reduce bias in labeled data used for the training and/or evaluation of machine learning models as well as to fund the development and maintenance of equitably labeled datasets most effectively and efficiently.”

‍

Namibia’s rangelands and ecosystems are under serious threat, their erosion and level of degradation are posing a risk on local biodiversity and food security. Overgrazing, frequent and prolonged droughts as well as lack of adequate land management resources are among the culprits. Moreover, climate change and bush encroachment decisively worsen the problem.

This pioneer project developed by Farm4Trade Namibia, with the support of the University of Namibia (UNAM) and other local partners, aims to create a new dataset that can lead and support the development of forecasting Machine Learning models to improve pasture management strategies and simplify decisions on the carrying capacity of selected areas.

The satellite imagery will be paired with their respective approximate land-cover labels. Remote sensing imagery will be annotated in collaboration with the local experts who will conduct the field activities on a monthly basis to observe the type of vegetation present in the selected areas. The collected data will include the vegetation stage, quantity estimates of perennial and seasonal plants, vegetation type, dominant plant species and the degree of grazing or browsing pressure. Data on livestock density in the study area will also be collected and associated with the labeling of satellite imagery. 

The project aims to bring benefits to multiple stakeholders of the agricultural and livestock sector and in particular to:

• Farmers participating in the data collection will benefit directly as they will receive reports and advice on rangeland practices and livestock production both during and after the project. They will also receive training on species identification and livestock stockage.
• Universities and MSc students in Rangeland Management of the University of Namibia will be involved in the project. In addition, students and supervisors can use these data for scientific publicatiotion, staff and students will also be trained and this will allow them to transfer the skills to future students or in their future workplaces. 
• Institutions, Farmers Unions and Associations will have access to the collected data having the possibility to share it to the communal farmers association, commercial farmers association thus ensuring the spread of data among their members and other institutions that work directly with farmers.

The creation of the aforementioned dataset will also allow Farm4Trade to apply its AI technologies to develop an innovative pasture management application able to automatically forecast grazing and pasture quantity and estimate the carrying capacity of a selected area. Overall, the data will empower the community to better manage, plan and mitigate rangeland health.

‍

‍

For more info about project funder visit lacunafund.org
For more info about the company visit farm4trade.com