About the Project
This project will use earth observation data combined with local ground truth from surveys to improve estimates of livestock diets in East Africa to support more robust estimates of Green House Gas (GHG) emissions from the livestock sector. Livestock are a key source of GHG emissions but the extent of their contribution is contested. Current numbers rely heavily on accurate estimates of livestock productivity and diet composition and these estimates are uncertain. For example, the FAO Global Livestock Environmental Assessment Model (GLEAM) uses crop distribution and yield data from GAEZ, in combination with feed baskets estimated by system and country, mostly derived from ‘expert opinion’. As well as uncertainty in composition, GLEAM is currently unable to account for seasonality in feed availability. Improvement of the accuracy of feed basket estimation and accounting for seasonality would significantly enhance the ability of the model to provide realistic estimates of GHG emissions from the livestock sector, particularly in LMICs, and to better estimate the mitigation potential of different interventions.
In this project we will combine expertise on livestock nutrition, GHG emissions estimation, modelling, machine learning and satellite image analysis to support a student to improve estimates of what livestock consume in East Africa, and how that impacts emissions from enteric fermentation and manure. The project will build on existing work to estimate feed availability in East Africa using publicly available land cover and dry matter productivity databases. The student will extend this work by exploring the use of machine learning algorithms for classifying very-high spatial resolution satellite imagery and spatial data such as primary household surveys and interrogation of existing secondary data on livestock diets in East Africa. Estimates of feed baskets will involve developing predictions of the various diet components in livestock diets based on cropping patterns, livestock density and type, intensification level, distance to market and a range of other metrics to be developed. The improved estimates will be incorporated into GLEAM on an experimental basis with the possibility to use the newly developed approaches for a wholesale revision of the feed basket component of GLEAM.
The student will be supervised by a multidisciplinary team drawn from The University of Edinburgh’s School of Geosciences, Royal (Dick) School of Vetinary Studies and the Global Academy of Agriculture and Food Security as well as the Food and Agriculture Organisation of the United Nations (FAO) and the International Livestock Institute (ILRI).
Research questions
1. Can livestock feed composition be accurately predicted using a range of metrics derived from household surveys and secondary data?
2. Does the use of Very-high spatial resolution satellite data (<1m spatial resolution) and machine learning enable identification of feed areas?
3. Can these predictions be used to develop spatial layers of livestock feed composition?
4. Do these new estimates improve the accuracy of GHG emissions estimates from the livestock sector?
The project will be split into a series of components as follows:
During the first year the student will assemble secondary data on livestock feed composition in East Africa using datasets available from within the supervisory team and their extended network of partners and collaborators and publicly available data. A literature review will be used to identify appropriate datasets and methods for the project and for the student to adapt the research questions to these literature findings and their own interests. Familiarisation with the GLEAM model and the FAO will also be encouraged within the first year.
During the second year the student will construct predictions of livestock feed composition, including seasonal variability, based on a series of hypotheses about what influences livestock diets, which will be developed with the expert supervision team and the literature review outcomes. These predictions will subsequently be applied to generate seasonal feed baskets for selected locations in East Africa.
During the latter stages of year 2 and into year 3 the predictions will be ground truthed by comparing them to the feed baskets using primary household survey data. Year three will culminate in experiments with incorporating improved feed baskets into GLEAM.
References
Havlík P, Valin H, Herrero M, Obersteiner M, Schmid E, Rufino M, Mosnier A, Thornton PK, Böttcher H, Conant RT, Frank S, Fritz S, Fuss S, Kraxner F, Notenbaert A. 2014. Climate change mitigation through livestock system transitions. PNAS 111(10):3709-3714.
Balehegn, M, Duncan, A, Tolera, A, Ayantunde, AA, Issa, S, Karimou, M, Zampaligré, N, Andre , K, Gnanda, I, Varijakshapanicker, P, Kebreab, E, Dubeux, J, Boote, K, Minta, M, Feyissa, F & Adesogan , A 2020, 'Improving adoption of technologies and interventions for increasing supply of quality feed in low- and middle-income countries', Global Food Security. https://doi.org/10.1016/j.gfs.2020.100372
Watmough, GR, Marcinko, CLJ, Sullivan, C, Tschirhart, K, Mutuo, PK, Palm, CA & Svenning, J 2019, 'Socioecologically informed use of remote sensing data to predict rural household poverty', Proceedings of the National Academy of Sciences (PNAS), pp. 201812969. https://doi.org/10.1073/pnas.1812969116
FAO (2019) GLOBAL LIVESTOCK ENVIRONMENTAL ASSESSMENT MODEL Version 2.0, Food and Agriculture Organisation of the United Nations, Rome http://www.fao.org/fileadmin/user_upload/gleam/docs/GLEAM_2.0_Model_description.pdf
FAO (2019) Five practical actions towards low-carbon livestock. The Food and Agriculture Organisation of the United Nations (FAO), Rome http://www.fao.org/3/ca7089en/ca7089en.pdf