M.Sc., Faculty of Science, University of the Witwatersrand, 2011 / Prolonged snow cover in Lesotho frequently results in human and livestock deaths, due to isolation and exposure. MODIS Rapid Response imagery is emerging as an important source of near real-time data for global hazard mapping and emergency response. A dataset of daily MODIS snow cover images for the period 2003 – 2010 was acquired for Lesotho. Combined with high-resolution SPOT satellite images for two study areas, MODIS snow cover images were used to establish the frequency, extent and timing of snowfalls. A digital elevation model was used in conjunction with mean air temperature data to investigate the effects of altitude, aspect and temperature on the mean rate of daily snowmelt. A strong correlation exists between mean day-time temperatures and the mean rate of daily snowmelt throughout the winter season. The mean rate of snow dissipation is most rapid after late season (September – November) snowfalls and least rapid after mid season (July – August) snowfalls. Snow cover persisting for 1 – 5 days dissipates at a higher mean rate than snow cover that has persisted for 6 – 10 days. Snow lasts longest on south-facing slopes above 2500m a.s.l, with evidence of increased ablation due to wind deflation and higher insolation levels in the highlands above 3400m a.s.l. The southern Drakensberg highlands in the district of Quthing have the highest mean duration of snow cover (21 – 25 days per annum). The seasonal extent and duration of snow cover was related to the spatial location of villages and roads in Lesotho, in order to determine individual vulnerability to negative impacts associated with prolonged snow cover. A ranking system was applied to each village according to the seasonal duration of snow cover, and the accessibility and proximity to the nearest road. Snowfalls occur between 1 and 8 times per annum on average. Therefore, village vulnerability is generally low, as most settlements are situated on predominantly north-facing slopes in the western lowlands and Senqu River Valley, which remain largely snow-free throughout the winter season. Few villages experience prolonged snow cover, which is limited to predominantly south-facing slopes above 2500m a.s.l along the escarpment and interior mountain ranges. Village vulnerability increases during the mid season period as a result of the increased frequency and duration of snow cover in July and August. The villages of Thoteng (Butha-Buthe), Letseng-la-Terae (Mokhotlong) and Mabalane (Butha-Buthe) have the highest vulnerability for the 2003 – 2010 period.

The emplacement of continental flood basalt provinces is often ascribed to fissure eruption. However, in many provinces the locations of actual vent complexes remains illusive. In southern Africa, the Karoo continental flood basalt province was erupted during the Jurassic between 183 and 179 Ma. The southernmost outcrop of the Karoo continental flood basalt lavas is the Lesotho remnant; and is comprised of the Drakensberg Group which forms the uppermost portion of the Karoo Supergroup. The geochemical stratigraphy for the Drakensberg Group is well established. At the base, there are a number of small volume compositionally diverse units, which form the Barkly East Formation, overlying which, there are larger volume, compositionally less variable units of the Lesotho Formation, which form the bulk of the volcanic sequence. The Lesotho remnant is associated with an abundance of dykes in the adjacent vicinity. This suggests that the lavas were fed from local rather than distal eruption sites. This study presents whole rock major and trace element data for 94 dykes and three sills from the northern Lesotho - northeastern Free State region and demonstrates that on the basis of geochemistry, all 97 intrusions can be correlated with various units of the northern Barkly East and Lesotho Formations. In addition, the petrographical; characteristics, orientation and distribution of the dykes do not correlate with geochemistry. Geochemical discrimination diagrams have been used to identify five compositionally diverse dykes, which are similar to the northern Barkly East Formation units. Three dykes are characteristic of the Letele unit and two are compositionally similar to the Wonderkop unit. Although the geochemical characteristics of the Lesotho Formation units are rather well constrained, the composition of these units is typified by considerable overlap in composition. Therefore, since unambiguous classification of dykes with geochemical similarities to the various units of the Lesotho Formation is unachievable using an empirical approach, the multivariate forward-stepwise discriminant function analysis (DFA) technique was used to facilitate the classification of the remaining 89 dykes and three sills. Forward-stepwise DFA classified 23 dykes as having compositional similarities to the Mafika Lisiu unit, 29 as having compositions of the Maloti or Senqu types; and 32 as having the composition of the Mothae type. In addition, eight dykes are compositionally similar to the Oxbow dykes, which intrude the Senqu unit in northern Lesotho. These results suggest that that the Lesotho remnant was fed from local eruption sites and that long distance magma transport for the bulk of the Lesotho remnant basalt lavas is unlikely.