Establishment of Trees for Sand Settlement in a Completely Desertified Environment

Abstract

In a desertified source area of drifting sand encroaching on parts of the Gezira Irrigation Scheme in Central Sudan, sand settlement by trees was envisaged. Establishment, survival and growth of Acacia tortilis, Leptadenia pyrotechnica, Prosopis juliflora, Salvadora persica and Panicum turgidum were investigated. Comparisons were made of the sums of final survivors in the replicates for each treatment as well as by an ANOVA analysis on the survival trends over time within the measuring periods. In a first experiment Acacia had better final survival but statistically comparable survival means with Prosopis over the year. This was due to higher deaths at later growth stages in Prosopis. Unprotected soil had some rain water run-on, while protection was not very effective, but irrigation generally was, particularly at later stages of growth. Establishing trees without irrigation appeared virtually impossible in a second experiment. Continuous drip irrigation under protection was very effective, but run-on water again influenced the protection issue. Leptadenia suffered from root hairs blocking water tubes. The higher survival rates of Acacia may be attributed to a capability of this species to grow on blown out eroded sites, on alkali soil. Differences in variation of soil conditions with depth on a small scale could explain growth contrasts in Acacia and Prosopis. An ANOVA on physical and chemical properties did not show such differences for Leptadenia, where soil properties causing a saturated zone possibly caused growth difference in a third experiment. Therefore irrigated unprotected Leptadenia with only weekly water topping-up did now better than protected trees. Leptadenia did much better in final surviving than Acacia, summed up for all conditions, and this was confirmed in the ANOVA results for throughout the year. Conclusions on design consequences for revegetation of such a completely desertified environment are drawn from the results.