Integrating indigenous and scientific knowledge systems for rangeland assessment in Somali region, Ethiopia

This study is part of Jigjiga One Health Initiative (JOHI), a project that aims at improving the health and livelihood of pastoralists in the Somali region of eastern Ethiopia. The health and livelihood of Somali pastoralists is inextricably linked with the “health” of the rangeland resources. Rangeland degradation, a decline in the “health” of rangeland resources, is one of the major challenges in the study area. Like other complex and contested problems, addressing rangeland degradation requires collaboration across sectoral, disciplinary and knowledge systems divides. Pastoralists, researchers, development practitioners and policy makers rely on indigenous and scientific knowledge systems to manage rangeland resources, guide rangeland researches, back rangeland interventions, and inform rangeland policies. For meaningful collaboration of all stakeholders, academic and non-academic actors need to have shared understanding on the status of rangeland resources and oregarding interventions needed to rehabilitate rangeland resources. Connecting the indigenous and scientific knowledge systems contributes to realization of a meaningful collaboration and to the establishment of integrated rangeland monitoring system needed for creation of shared understanding. The aim of this thesis was to connect the indigenous and scientific knowledge for integrated rangeland resources assessment. The objectives of this study were to study environmental differences and changes that matter for pastoralists and their perception on drivers of environmental changes; to explore if perceived environmental differences, environmental changes and drivers of environmental changes are detectable using moderate resolution satellite images; and to examine where the indigenous and scientific knowledge systems matches and mismatches. The study was undertaken in Adadle woreda, Somali region, Eastern Ethiopia, using both qualitative and quantitative data. Qualitative data were gathered using focus groups discussion (FGD), key informant interviews and field observations. The discussions and interviews were conducted in Somali language and were recorded for later transcription. Guiding questions were used for interviews that focused mainly on traditional rangeland management; rangeland classification; rangeland assessment and monitoring; and pastoralists’ perception on rangeland condition, on social and ecological drivers of environmental changes and their consequences. All audios of FGD and KII were transcribed and translated into English and analyzed using ATLAS.ti 8. Both a priori and emrgent codes were used for coding the qualitative data. Priori codes were based on research objectives and researchers experience with pastoralists; emergent codes were developed from new and unanticipated themes that emerged from interviews and discussions. Excerpts of the translated interviews were included to support the analysis and illustrate how pastoralists express the environmental differences and changes that matter to them. For quantitative data, CHIRPS, ERAS ECMWF, Landast SR and SRTM data sets from Google Earth Engine was used. The quantitative data analysis comprised trend analsis of climatological data (precipitation and tempreture), NDVI, NDWI and land cover analysis. Herders in the study area divide their territory into two macro landscapes: the upland and lowland macro-landscapes. They prefer the upland macro-landscape due to higher abundance of palatable species and lower livestock disease risk. Macro-landscapes are further divided into micro-landscapes named after after dominant fodder species, wildlife, historical events and topographic feattures. Based on the perceived quality of micro-landscapes for livestock production, herders categorized their territory into “Dal-san” (good quality land) and “Dal-xun” (poor-quality land). Indigenous rangeland classification influences how herders use and monitor rangeland resources in the study area. The upland macro-landscape serve as wet season grazing area and drought reserve, while the lowland macro-landscape is used as dry season grazing area. Pastoralists in the study area use both ecological (e.g. vegetation structure and composition, palatability) and anthropogenic (e.g. livestock product and reproductive performance) indicators to assess environmental changes. According to herders, the abundance of palatable species has declined over the last three decades, while the abundance of unpalatable and poisonous species has increased in the same period. Besides changes in species composition, the structure of the rangeland vegetation has changed, grasslands either encroached by woody species or replaced with annual species of lesser value, or converted into barren land. In line with herders’ assertion of higher abundance of fodder in the upland macro-landscapes, NDVI comparison between the upland and lowland macro-landscapes during rainy season showed slightly higher NDVI values for the upland macro-landscapes. Contrary to herders claim, there was no significant difference of mean NDVI between “good-quality” and “poor-quality” micro-landscapes over the period of 1988-2020. Land cover change analysis over three decades (between 1990 and 2020) showed varying changes among the different land cover classes. According to herders, the drivers of environmental changes in the study area include decline in rainfall amount and duration; deforestation; and loss of land to competing land uses (e.g. private rangeland enclosures, irrigated farming and settlements); increase in livestock population; declining livestock mobility; establishment of Kebeles; opening of new water points; weakening of traditional institutions; and poverty. In line with herders claim, analysis of settlement expansion of five Kebeles in Adadle woreda between 2010 and 2020 showed significant change in the spatial extent of settlements. NDVI analysis of settlement also confirmed negative impact of expanding settlement on nearby environment. Despite contrary claims by herders, analysis of rainfall data of the main rainy season did not show a decline in rainfall over the prior three decades. To conclude the indigenous rangeland classification and landscape labelling of pastoralists influence how they monitor and use rangeland resources. Pastoralists focus on environmental changes that affect their livelihood (the decline of key fodder species or increase in unpalatable or poisonous species). Pastoralists attribute environmental changes to various drivers including rising demographic changes, increasing sedentarization of pastoralists, worsening poverty and declining rainfall. The mismatch between herders’ perception of environmental changes and absence of significant differences could be due to limitation of moderate resolution satellite images used for this study and/or conceptual differences between herders understanding of rainfall and rainfall measured by scientific tools. Development practitioners and policy makers need to consider herders quality-based landscape classification and their livestock species preference in prioritization of interventions sites for rangeland rehabilitation. We recommend the development of integrated rangeland resources assessment. In the integrated rangeland assessment system, pastoralists contribute their indigenous rangeland classification and labelling, their undertstanding of lanscape suitability and risks of livestock disease and conflict. Researchers undertake NDVI and rainfall anomally analysis using Google Earth Engine. By connecting information and knowledge generated by the two knowledge system, a system to forcast available fodder can be developed and used for co-management of rangeland resources and for early warning system. By creating a possibility of linking pixels with people living in the pixels, joint use of Google earth engine and mobile technologies makes connecting the two knowledge systems plausibl