Recovery of vegetation and bees after removal of pine forests by fire in the Limietberg region of the Cape Floristic Region biodiversity hotspot

[摘要] ENGLISH ABSTRACT: The present global biodiversity crisis is characterized mostly by loss of species due to habitatdestruction but there other major threats – notably invasive alien species. The term 'biodiversity hotspothas been coined to emphasize areas for conservation prioritization – areas that have high biodiversity underthreat from habitat destruction, invasive species, etc. The Cape Floristic Region (CFR) in the Western Cape(WC) Province of South Africa has the highest biodiversity of all Mediterranean-type climate regionsglobally and is classified among global biodiversity hotspots. The CFR, a Mediterranean-type climateheathland with a naturally high fire frequency, faces threat from transformation by development, disturbed(increased) fire frequencies and intensities, and invasive organisms, most notably invasive trees. Suchthreats disturb keystone species and keystone processes including insect pollinator assemblages andassociated insect flower visitation webs and frequencies.Invasive pines are a serious threat to insect flower visitation as pine trees (Pinus spp.) shade outmuch indigenous vegetation in the CFR. Little is known of their long-term effects on insect flower visitorsand vegetation recovery in post-pine restoration and recovery areas after such trees are removed. Iinvestigated the recovery of vegetation and the most important insect pollinator, bees, after the removal ofpines by fire and passive recovery in a CFR valley in the Western Cape.In 1999, a wildfire burned much of the WC Limietberg Nature Reserve along with an adjacent pinetree forestry stand - which was then left to recover, providing an ideal opportunity to investigate the enduringeffects of pine afforestation in the CFR. In two data chapters, I compared the post-fire passive recovery ofvegetation (Chapter 2) and bee diversity (Chapter 4) in areas which had previously been planted with pinesvs. those which had contained natural fynbos. To improve on sampling methodology, I conducted anexperiment that demonstrated the Observer Effect in bee sampling with a sweep net (Chapter 3), and Ideveloped a novel sampling device (Chapter 5) for insect flower visitors.Sampling of vegetation and bee diversity was conducted in a paired sampling design, where fynbos(Natural) sub-sites were paired with sub-sites which had previously been afforested with pines (Post-PineRecovery; PPR) and the two sub-sites were separated by a distinct, linear boundary (Natural/PPR boundary).Sampling was conducted along transects parallel to the boundary and extending in both directions from theboundary into the Natural and in the opposite direction into PPR sub-sites. Five transects were positioned at3, 10, 20, 30, and 40 m from the boundary (Ecotone) and three were positioned at 60, 80, and 100 m from theboundary (Deep).In Chapter 2, I found that natural sub-sites consistently had higher total plant species abundance andspecies richness than PPR sub-sites. Approximately two thirds of plant species were more abundant inNatural sub-sites than in PPR sub-sites. There was no significant correlation in species richness orabundance with distance from the Natural/PPR boundary. Some genera are cautiously indicated as having lower success in recovery after pine afforestation: Erica spp., Restio spp., Hypodiscus spp. whileHelichrysum spp. is also tentatively indicated to recover well in PPR sub-sites. Soil disturbance andconcomitant disruption of normal ecosystem functions, including pollination, is indicated as a probablereason for disruption of plant recovery.In Chapter 3, sweep netting methodology was tested for the Observer Effect. A noticeable increasein bee visitation frequency to a common generalist plant species in bloom, Metalasia densa, was correlatedwith longer waiting periods after I stopped moving indicating the presence of an Observer Effect. Thissuggests that sweep netting for bees should only be commenced after a waiting period of five minutes duringwhich the sweep netter is motionless.In Chapter 4, using a flight-intercept modified pan trap, I compared bee species richness and beespecies abundance across different seasons and in both mass-flowering and sparser flower patches. Therewas no significant difference in bee species richness between Natural and PPR sub-sites. All bee species,except one complex, were more abundant in Natural sub-sites. Nearly two thirds of all bee species (n=37 of56) caught with sweep netting and the modified pan trap are undescribed species. Similar to the vegetation,the effects of soil disturbance as a result of decades of pine forest shading and pine forest litterfall followedby an unusually hot pine forest fire are indicated as the most likely reasons for lower bee abundance in PPRsub-sites. This is due to the difficulty associated with viable nest establishment and suitable pollen andnectar forage availability in disturbed areas.In Chapter 5, the newly developed Pan and Flight Intercept Combination (PAFIC) trap's design,pilot testing, and comparison with the traditional pan trap is discussed. A preliminary test suggests that thePAFIC trap is more efficient (with higher abundance) than the traditional pan trap.In Chapter 6, I discuss the implications of the disturbance of pine forestry and unusual pine fire toplant species and bee species assemblages. Bee-pollination webs in PPR sub-sites are indicated as beingsubstantially simpler than those in Natural areas as well as compositionally different. The recovery ofpollination as a keystone process in post pine-afforestation areas faces a substantial challenge given thedisturbance to soil that decades of pine afforestation followed by pine forest fire can cause. Somesuggestions are made for the restoration of fynbos areas recovering from pine afforestation including adiscussion of augmenting re-vegetation measures (fynbos seed dispersal and seedling planting) with methodsof restoring of healthy pedogenesis, epigaeic arthropod communities, and fynbos seedbanks.

[发布机构] Stellenbosch University