Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Titre : SOIL GREENHOUSE GAS EMISSIONS FROM DOMINANT LAND USE TYPES IN LOWLANDS AND HIGHLANDS OF SOUTHERN KENYA

Auteur : Wachiye, Sheila

Université de soutenance : University of Helsinki,

Grade : Doctoral dissertation (article-based) / Doctoral Programme in Atmospheric Sciences 2022

Résumé
Land-use changes account for three-quarters of Africa’s anthropogenic greenhouse gas (GHG) emissions, yet little is known about these emissions from a wide variety of land uses, especially from the savanna in Sub-Saharan Africa. To address this gap, I conducted four studies (2018 - 2021) to quantify soil GHG emissions from dominant land use types (LUTs) in the Taita Taveta county, southern Kenya. Three studies were carried out in the lowland under savanna grassland (private conservation and community grazing), bushland, maize cropland (Zea mays L.), and sisal plantation. The fourth study was in the highlands under indigenous forest, pine and eucalyptus plantations, woodland, and maize cropland. Each study covered a whole year. Soil GHG concentrations were measured with static GHG chambers and analysed by gas chromatography over eight one-week and weekly campaigns. Environmental factors including soil temperature, volumetric soil moisture content, vegetation, soil physicochemical characteristics and grazing intensity were evaluated. Daily CO2 emissions varied between 5 – 360 mg C m-2 h-1 and differed between sites. In the lowland, grassland under conservation land had the highest CO2 emissions, while cropland and bushland had the lowest. In the highlands, the pine plantation and indigenous forest emitted higher CO2 than eucalyptus and woodland, while cropland emitted the least. Overall, soil organic carbon accounted for 50% variation in CO2 emissions between the LUTs, while grazing accounted for 60% variation between grassland under conservation and the community grazing. In the sisal plantation, the oldest and youngest stands emitted higher CO2 than mature stands, which is consistent with root respiration since soil organic carbon did not differ between stands. During each study year, CO2 emissions were highest in the wet than in the dry season due to the higher soil moisture and vegetation cover. In the lowland, soil moisture alone explained 20 – 70% seasonal variation in CO2 emissions, while vegetation (based on EVI and NDVI ; proxies of vegetation status) explained 18 – 82%. The combined influence of soil moisture and vegetation dynamics proved to be a greater control accounting for 75 – 96% of seasonal variation in CO2 emissions. The indigenous forest and pine plantation were affected by soil temperature rather than moisture, but the relationship was very weak (< 20%). In the lowland, soil N2O (-7.7 – 17 µg N m- 2 h-1) and CH4 (-0.4 – 1.9 mg C m-2 h-1) emissions were very low throughout each study period. Most CH4 values were below the limit of detection (LOD ; ±0.03 mg C m-2 h-1) and the ones above were sinks rather than sources, and they did not differ between seasons or sites. N2O emissions were highest in cropland and lowest in bushland but did not vary with season. In the maize cropland, manure application was very low and had no effect on N2O emissions. Due to low soil carbon, nitrogen, and moisture content at all LUTs, N2O and CH4 emissions were of minor significance. Among the assessed LUTs, soil carbon content, moisture, and vegetation characteristics were key drivers for soil GHG emissions. The findings of this study address a significant gap in the spatial and temporal variability of soil GHG emissions in Sub-Saharan Africa, which can be used to improve current estimates, thereby improving climate change mitigation measures when identifying and quantifying national contributions.

Présentation

Version intégrale (3 Mb)