Effects of nitrogen fertilization and precipitation decline on tree-grass interactions in Southern African Savannas

Abstract

Savannas are heterogeneous ecosystems comprising of a co-dominance of herbaceous and woody species and these ecosystems support millions of people as well as diverse mammalia species. Recent research has established that savannas are undergoing structural changes mainly characterized by an increase in the woody component commonly referred to as woody encroachment. However, this widespread phenomenon is still poorly understood. This study sought to fill this gap by using innovative wavelet based methods to test the existence of this phenomenon as well as quantify its extent. Using a four year field experiment and satellite data the study tested whether and to what extent woody encroachment is explained by variations in key resources, that is, moisture availability and nitrogen. In the experiment, rainfall was reduced by 15% using fixed location rainout shelters while Nitrogen fertilizer was added as ammonium nitrate (NH4NO3) at a rate of 4 grams of N m-2 at the start of the rainfall season each year. As rainfall received at the study site varied across the study period, the amount of rainfall suppressed was variable with as little as 50 mm (for a rainfall total = 337 mm) in 2012 to as much as 125 mm in 2014 (for a rainfall total 832 mm). Results of wavelet analysis on multi-temporal imagery point towards an increase in woody encroachment thereby providing new insights into the phenomenon. For instance, at Lake Chivero where woody patches were already established the dominant scale of woody patches quadrupled over the 39-year period while the intensity decreased slightly by 0.7%. In contrast, at Kyle which was predominantly a grassland, the dominant scale decreased from 32 m to 16 m over the same period while intensity increased by >200%. Results further indicate that decadal changes in rainfall significantly explain woody patch dynamics in semi-arid savanna landscapes. In fact, decadal changes in rainfall significantly explained more than 85% of the ii variation in woody patch dynamics (R2 = 0.85, F13 = 35.9, P < 0.01). In addition, results from field experiments indicate that moisture availability and its interaction with nitrogen had significant (P<0.05) but transient effects on woody species recruitment in a semi-arid savanna ecosystem using Vachellia karroo as a focal species. Though not significant, woody saplings growing under suppressed rainfall were, overall, 9% shorter (64.72±13.98cm SD) than the control (71.034±13.982cm SD). In contrast, saplings growing under rainfall suppression had a slightly larger diameter (1.153±0.348cm SD) than those growing in the control (1.056±0.418cm SD). However, grass competition did not significantly influence the growth of V. karroo species either alone or in interaction with nitrogen and rainfall. Future climate change projections indicate the bioclimatically suitable range of V. karroo in southern Africa will expand despite the projected rainfall decrease. Results of this thesis further the understanding of woody encroachment and its key drivers in a changing savanna environment.