Thesis (PhD(For))--Stellenbosch University, 2012. / The subtropical ecosystem of the Zululand coastal plain is prized by the South African commercial plantation forestry industry for its rapid clonal Eucalyptus growth, short rotations (6 to 7 years) and high yields. This region is typified by sandy soils that are low in clay and organic matter, have small nutrient reserves and are poorly buffered against nutrient loss. The subtropical climate induces rapid decomposition of residues and tree litter resulting in small litter nutrient pools and rapid nutrient release into the soil, particularly after clearfelling. A combination of large nutrient demands through rapid growth, rapid nutrient turnover and small soil nutrient reserves implies that sites in this region are sensitive and may be at risk of nutrient decline under intensive management. The work in this study set out to determine the risk of nutrient depletion through harvesting and residue management on a site within the Zululand region, to assess nutritional sustainability and the risk of yield decline in successive rotations. Some bulk biogeochemical cycling processes of macro-nutrients nitrogen (N), phosphorus (P), potassium (K), calcium (Ca) and magnesium (Mg) were assessed, and assessments also included sodium (Na). An existing Eucalyptus stand was clearfelled and treatments were imposed on the residues after broadcasting to simulate various levels of nutrient loss through levels of harvesting intensity and residue management. These included residue burning (Burn), residue retention (No-Burn), fertilisation (stem wood nutrient replacement), whole tree harvesting and residue doubling. Outer blocks of the stand were not felled, but included as replicates of an undisturbed standing crop treatment. Biogeochemical nutrient cycling processes were assessed primarily in the standing crop, Burn and No-Burn treatments, in the assumption that these represented the furthest extremes of nutrient loss. Data collection commenced a year prior to clearfelling and continued to two years and six months after planting with key data collection over a 20.1 month period from clearfelling to canopy closure (one year after planting). Water related nutrient pools and fluxes were assessed as atmospheric deposition (bulk rainfall, throughfall and stemflow) and gravitational leaching to 1m soil depth. Drainage fluxes were predicted using the Hydrus model and real-time soil moisture data. Zero tension lysimeters collected soil solution for chemical analysis. Sequential coring in the 0 to 30cm soil layer was used to determine in situ soil N mineralisation. Soil chemical and physical properties were assessed over the first meter of soil at clearfelling and new crop canopy closure to determine soil nutrient pools sizes. Biomass nutrient fluxes were assessed from litterfall, residue and litter decomposition, and above ground accretion into the tree biomass. Leaching and N mineralisation were monitored in the No-Burn, Burn and standing crop treatments only. Atmospheric deposition, while variable, was shown to be responsible for large quantities of nutrients added to the Eucalyptus stand. Nitrogen and K additions were relatively high, but within ranges reported in previous studies. Rapid tree canopy expansion and subsequent soil water utilisation in the standing crop permitted little water to drain beyond 1m resulting in small leaching losses despite a sandy well drained soil. Further leaching beyond this depth was unlikely under the conditions during the study period. Mineralisation and immobilisation of N also remained low with net immobilisation occurring. The standing crop was shown to be a relatively stable system that, outside of extreme climatic events, had a relatively balanced or positive nutrient budget (i.e. nutrient inputs minus outputs). Large quantities of nutrients were removed with stem-wood-only harvesting in the No-Burn treatment leaving substantial amounts on the soil surface in the harvest residues. Whole tree removal increased losses of all nutrients resulting in the largest losses of P and base cations compared to all other treatments. This was mostly due to high nutrient concentrations in the removed bark. Loss of N in the Burn treatment exceeded whole tree N losses through combustion of N held in the harvest residues and litter layer. The majority of K leached from the residues prior to burning and a relatively small fraction of the base cations were lost from the partially decomposed residues during burning. Ash containing substantial amounts of Ca and relatively large amounts of N and Mg remained after burning. Surface soil Ca and Mg was significantly increased by the ash which moved into the soil with rainfall directly after burning. Rapid soil moisture recharge occurred within a few months after clearfelling, increasing leaching from the upper 50cm of soil. Clearfelling increased net N mineralisation rates, increasing mobile NO3-N ions in the soil surface layers. Nitrate concentration peaked and K concentration dipped in the upper soil layers of the Burn treatment directly after burning. Deep drainage and leaching (beyond 1m depth) over the 20.1 month period was, however, not significantly different between the Burn and No-Burn treatments. Rapid soil moisture depletion and nutrient uptake with new crop growth reduced leaching fluxes to levels similar to the standing crop by six months after planting. Taking the full rotation into account, clearfelling induced a short-lived spike in N and cation leaching compared with the low leaching losses in the undisturbed standing crop. Soil N mineralisation over the 20.1 month period in the burnt treatment was half that of the No-Burn treatment. Growth and nutrient accumulation was significantly higher in the fertilised treatment than in other treatments up to 2.5 years of age. Growth in the Burn treatment was greatest compared to other treatments during the first few months, but slowed thereafter. No significant growth differences were found between all other treatments from a year to 2.5 years after planting. Early growth was therefore apparently not limited by N supply despite large differences in N mineralisation between Burn and No-Burn. Foliar vector analysis indicated that fertilisation improved growth initially through increased foliar N and P at six months after planting followed by Mg and Ca at one year. The Burn treatment was not nutrient limited. These growth results contrasted with similar international research on sandy tropical sites where growth was reduced after residue removal and increased after residue doubling. The combined nutrients released from pools in the litter layer or ash and soil in addition to atmospheric inputs were sufficient to provide most nutrients required to maintain similar growth rates across all treatments. This demonstrated the importance of residue derived nutrients to early growth nutrient supply. Reduced N mineralisation through a lack of substrate may limit N supply later in the rotation where residue had been removed. Construction of a nutrient budget for the system revealed that high levels of atmospheric inputs have the potential to partially replenish a large proportion N, K, and Ca lost during clearfelling, provided losses are constrained to stemwood removal only. However, loss of Mg that occurred primarily through leaching may not be replaced under the low Mg inputs recorded in this study. Larger nutrient removals (i.e. stemwood plus other plant parts) placed a heavier reliance on the small soil nutrient pools at this site which can limit future productivity. More intense harvesting and residue management practices dramatically increased the risk of nutrient depletion. Losses of specific nutrients depended on a combination of clearfelling biomass removal, residue burning and subsequent leaching. Nitrogen losses due to harvesting and burning were more substantial than those due to leaching. Mg and K losses depended most strongly on the time after clearfelling before re-establishment of the new crop and rainfall patterns, while Ca and P losses depended directly on the amount of biomass removed. Depletion risk was the greatest for Mg and K through rapid leaching, even after stem wood only removal. Deep root uptake and deep drainage with associated cation loss needs to be investigated further to quantify ecosystem losses and recovery of cations displaced beyond 1m. Atmospheric deposition is one of major factors countering nutrient losses. However, atmospheric inputs may not be reliable as these may lessen in future through pollution control legislation and climate change. Changes in growth rate under poor nutrient management practices are small and difficult to detect relative to the large impacts of changing weather patterns (drought), wildfire and pest and disease. This makes it difficult to prove nutrient related growth decline. It may be possible that improvements in genetics, silvicultural technologies and atmospheric inputs may also be masking site decline (in general) and in part explain the lack of evidence of a growth reduction in the region. As the poorly buffered sandy soils on the Zululand Coast are at risk of nutrient depletion under the short rotation, high productivity stands, it may be necessary to stipulate more conservative harvesting and residue management practices. A more conservative stem-wood only harvesting regime is recommended, retaining all residues on site. Residue burning should be avoided if N losses become a concern. The length of the inter-rotation period must be kept short to reduce cation leaching losses. Site nutrient pools need to be monitored and cations may eventually need to be replenished through application of fertilisers or ash residues from pulp mills. Management practices therefore need to be chosen based on the specific high risk nutrients in order to maintain a sustainable nutrient supply to current and future plantation grown Eucalyptus.