Fire-modulated fluctuations in nutrient availability stimulate biome-scale floristic turnover in time, and elevated species richness, in low-nutrient fynbos heathland

Abstract

Background and aims: In many systems, postfire vegetation recovery is characterized by temporal changes in plant species composition and richness. We attribute this to changes in resource availability with time since fire, with the magnitude of species turnover determined by the degree of resource limitation. Here, we test the hypothesis that postfire species turnover in South African fynbos heathland is powered by fire-modulated changes in nutrient availability, with the magnitude of turnover in nutrient-constrained fynbos being greater than in fertile renosterveld shrubland. We also test the hypothesis that floristic overlaps between fynbos and renosterveld are attributable to nutritional augmentation of fynbos soils immediately after fire.

Methods: We use vegetation survey data from two sites on the Cape Peninsula to compare changes in species richness and composition with time since fire.

Key results: Fynbos communities display a clear decline in species richness with time since fire, whereas no such decline is apparent in renosterveld. In fynbos, declining species richness is associated with declines in the richness of plant families having high foliar concentrations of nitrogen, phosphorus and potassium and possessing attributes that are nutritionally costly. In contrast, families that dominate late-succession fynbos possess adaptations for the acquisition and retention of sparse nutrients. At the family level, recently burnt fynbos is compositionally more similar to renosterveld than is mature fynbos.

Conclusions: Our data suggest that nutritionally driven species turnover contributes significantly to fynbos community richness. We propose that the extremely low baseline fertility of fynbos soils serves to lengthen the nutritional resource axis along which species can differentiate and coexist, thereby providing the opportunity for low-nutrient extremophiles to coexist spatially with species adapted to more fertile soil. This mechanism has the potential to operate in any resource-constrained system in which episodic disturbance affects resource availability.

Keywords: Alpha diversity; community richness; fire; fynbos; postfire succession; resource availability; soil nutrients; species richness; temporal turnover; vascular plants.

Fig. 1.

Relationships of species richness to postfire vegetation age across the Cape Point fynbos plots (A) and the Signal Hill renosterveld plots (B).

Fig. 2.

Relationships of species richness to postfire vegetation age across the Cape Point fynbos plots, by family. Only families present in ≥20 plots are included. Given that symbols have a transparent fill, dark points represent instances of multiple overlapping points. The significance of relationship is indicated by line thickness, with thin lines representing P < 0.05, medium lines P < 0.01 and thick lines P < 0.001.

Fig. 3.

Maximum likelihood phylogenetic reconstruction of the mean age of the vegetation occupied by the Cape Point species. Branch shading indicates vegetation age association, with red and blue, respectively, indicating an association with young (i.e. recently burnt) and old vegetation. Important family-level lineages are indicated on the right of the tree.

Fig. 4.

Relationships of species richness to postfire vegetation age across the Signal Hill renosterveld plots, by family. Only families present in ≥20 plots are included. Given that symbols have a transparent fill, dark points represent instances of multiple overlapping points. The significance of relationship is indicated by line thickness, with thin lines representing P < 0.05, medium lines P < 0.01 and thick lines P < 0.001.

Fig. 5.

Box-and-whisker comparison, by family, of the species richness–vegetation age correlation, for the Cape Point fynbos (Fig. 2) and Signal Hill renosterveld (Fig. 4) plots. The six families included, by virtue of their presence in ≥20 plots at both Cape Point and Signal Hill, are as follows: AS, Asteraceae; FA, Fabaceae; IR, Iridaceae; PO, Poaceae; RU, Rubiaceae; SC, Scrophulariaceae. The whiskers depict the minimum–maximum ranges, the boxes the 0.25–0.75 interquartile ranges, and the dark bar the median, across the six families.

Fig. 6.

Relationship of the species richness–vegetation age correlation to mean foliar nitrogen (A, B), phosphorus (C, D) and potassium (E, F) concentrations, across families present in ≥20 plots at Cape Point (left panels; n = 10) and Signal Hill (right panels; n = 6). Family codes are as follows: AS, Asteraceae; CY, Cyperaceae; ER, Ericaceae; FA, Fabaceae; IR, Iridaceae; PO, Poaceae; PR, Proteaceae; RE, Restionaceae; RU, Rutaceae; SC, Scrophulariaceae. Regression lines are fitted using both ordinary least squares (black lines) and phylogenetically generalized least squares (grey lines), with the outlying Iridaceae (IR) both included (solid lines) and excluded (dotted lines).

Fig. 7.

(A) Non-metric multidimensional scaling biplot extracted from a site × family community matrix comprising both the Cape Point fynbos (triangles) and Signal Hill renosterveld (circles) plot data. Cell entries in the input matrix describe the number of species representing a particular family at a particular site, and between-plot distances are based on Bray–Curtis dissimilarity. Symbol size describes postfire vegetation age, with larger symbols representing recently burnt plots. The associated stress value is 0.093. (B) Relationship of the MDS1 score (from panel A), describing the family-level compositional similarity of the Cape Point fynbos plots to the Signal Hill renosterveld plots (high values indicate greater similarity), as a function of postfire fynbos vegetation age.