INTERCONNECTION BETWEEN FEED RESOURCES, LIVESTOCK PRODUCTION AND SOIL CARBON DYNAMICS UNDER THE CHANGING CLIMATE IN EASTERN ETHIOPIA

Abstract

Livestock are still the major vehicle to support pastoral and agro-pastoral livelihoods in terms of food security and income generation in eastern Ethiopia. However, climate variability is negatively affecting the livestock systems through reducing water and feed resources. In this context, failure to timely understanding and responding to this environmental influence under smallholder conditions would lead to worst scenarios on livestock production and thereby affecting the livestock herders’ livelihoods. Therefore, the objective of this PhD dissertation was to understand the inter-connection between rainfall variability, livestock performance, feed resources availability and quality, as well as herd size and soil carbon balance in semiarid areas of eastern Ethiopia. Moreover, perception of livestock herders on challenges and their adaptation strategies under changing climate was also assessed. The data of long-term rainfall (1984-2015) were obtained from the National Meteorology Agency; livestock population and its production, mortality and off take rates from Central Statistics Agency (1992-2015), and respective livestock herders. El-Niño Southern Oscillation (ENSO) events information was also collected from the National Oceanic and Atmospheric Administration. Moreover, livestock herder’s perception, challenges and adaptation strategies posed by rainfall variability were obtained using semi-structured questionnaires from 384 household heads interviews. Rainfall trend was analyzed using Mann-Kendall’s statistical tests. Coefficients of variation (CV), and standardized rainfall anomaly (SRA) were used to calculate rainfall variability, and the drought frequency, respectively. The Livelihood Early Assessment and Protection (LEAP) software was used to calculate the rangeland water requirement satisfaction index (WRSI). In addition, regression analysis was used to assess the association of rainfall variability with livestock population, milk yield, mortality, off-take rate, and rangeland WRSI. Representative Concentration Pathway 4.5 scenario was used for determining rainfall projection. The JAVA model was used to study the relationship between feed availability, livestock performance and carbon balance. The Mann-Kendall trend test showed an insignificant increasing and decreasing trend of annual rainfall in all study zones and districts. In contrast, 92% of the respondents perceived a decreasing trend of annual and xvi seasonal rainfall in all study districts. The future rainfall trend is also expected to increase (P>0.05). The CV values for both current and future rainfall confirmed moderate (23.1%) to high (35.4%) variability of rainfall. The SRA indicated that drought has occurred every two to five years. The mean annual rainfall variability had strong association (P<0.05) with cattle and sheep population, and milk yield in pastoral and agro-pastoral production systems. Cattle and sheep mortality and off-take rates in pastoral and agro-pastoral production systems had also a significant negative association with rainfall variability. Moreover, there were a strong association between rangeland WRSI values and the corresponding rainfall amount in Jigjiga (P<0.001, r2 = 0.80) and Shinile (P<0.001, r2 = 0.87) pastoral areas. Similarly, short and long rainy seasons had strong associations (P<0.001) with rangeland WRSI. Furthermore, the predicted future mean annual rainfall scenario 2020-2049, 2040-2069 and 2070-2099 in all districts showed variability in distribution and would be further leading to reduction of rangeland WRSI. Livestock herders also perceived the impact of rainfall variability on livestock population and productivity and applied several adaptation strategies including settlement around watering points, seasonal mobility, and destocking. However, lack of timely climate information and mobility restriction were major adaptation challenges under pastoral communities. The chemical composition, digestibility and degradability of browse trees were affected by season. Almost all browse species had a moderate to high crude protein (CP) content (52.4 - 220 g/kg DM), and a high IVDMD (623 - 678 g/kg DM) and metabolisable energy values (8.80 - 9.70 MJ/kg DM) during the wet season. Fiber (NDF, ADF, ADL, hemi-cellulose, cellulose), ash, phenol and tannin contents were higher during dry season than wet season. The JAVA model showed that mean daily live weight gain and milk production per TLU continuously increased with decreasing herd size (HS). The total annual live weight gain reached a maximum at the use of the 40% best feeds and a HS of 722 TLU. However, soil carbon balance is negative and deteriorated with increasing feed use. In conclusion, provision of early warning, timely weather forecast information, and adjusting actual livestock population to the level of feed availability are critical to increase livestock production and enhance adaptive capacity of livestock herders in the study areas.