Systematic reduction of natural enemies and competition across variable precipitation approximates buffelgrass invasiveness (Cenchrus ciliaris) in its native range

Abstract

Invasive grasses cause devastating losses to biodiversity and ecosystem function directly and indirectly by altering ecosystem processes. Escape from natural enemies, plant-plant competition, and variable resource availability provide frameworks for understanding invasion. However, we lack a clear understanding of how natural stressors interact in their native range to regulate invasiveness. In this study, we reduced diverse guilds of natural enemies and plant competitors of the highly invasive buffelgrass across a precipitation gradient throughout major climatic shifts in Laikipia, Kenya. To do this, we used a long-term ungulate exclosure experiment design across a precipitation gradient with nested treatments that (1) reduced plant competition through clipping, (2) reduced insects through systemic insecticide, and (3) reduced fungal associates through fungicide application. Additionally, we measured the interaction of ungulates on two stem-boring insect species feeding on buffelgrass. Finally, we measured a multiyear smut fungus outbreak. Our findings suggest that buffelgrass exhibits invasive qualities when released from a diverse group of natural stressors in its native range. We show natural enemies interact with precipitation to alter buffelgrass productivity patterns. In addition, interspecific plant competition decreased the basal area of buffelgrass, suggesting that biotic resistance mediates buffelgrass dominance in the home range. Surprisingly, systemic insecticides and fungicides did not impact buffelgrass production or reproduction, perhaps because other guilds filled the niche space in these highly diverse systems. For example, in the absence of ungulates, we showed an increase in host-specific stem-galling insects, where these insects compensated for reduced ungulate use. Finally, we documented a smut outbreak in 2020 and 2021, corresponding to highly variable precipitation patterns caused by a shifting Indian Ocean Dipole. In conclusion, we observed how reducing natural enemies and competitors and certain interactions increased properties related to buffelgrass invasiveness.

Keywords: insect herbivory; invasiveness; natural enemies; pathogens; plant competition; ungulate herbivory.

FIGURE 1

Buffelgrass associates: Top left: Buffelgrass seed head with fungal smut infection. Top right: healthy stand of buffelgrass. Bottom left: Stem gall exit holes generated by Tetramesa sp. stem‐galling wasp (L) and other insect (probably a beetle) (R). Bottom center: Galls with exit holes made by Orseolia sp. stem‐galling midges. Bottom right: Elephant grazing on grass.

FIGURE 2

(a) Mpala Research Centre has six study locations along a precipitation gradient of 400 to 600 mm. (b) Mpala Research Centre shown in an inset map of Kenya's political boundaries. (c) A fenced and an unfenced plot are shown adjacent. Each gray box above the dotted line represents one of six randomized blocks across the 50 m by 50 m plot. (d) A zoomed‐in depiction of the randomized block with the four treatment subplots labeled as F – Fungicide, N – Control, I – Insecticide, and CR – competition release. Subplots were randomly arranged and spaced within each block.

FIGURE 3

(a) Buffelgrass basal area as a function of fenced and unfenced plots is shown across the study years. (b) The effect of the treatments is displayed on the same scale as panel (a). The error bars for each panel represent one standard error.

FIGURE 4

(a) Buffelgrass height is shown as a function of fenced and unfenced plots across the study years. (b) The effect of the treatments is displayed on the same scale as panel (a). The error bars for each panel represent one standard error.

FIGURE 5

Interaction plots showing the model estimated height at the low lag precipitation (100 mm) and high lag precipitation (560 mm) range. Fenced plots are denoted with a solid line and filled circles, while the unfenced plots are denoted with a dashed line and open circles. The standard error is shown with the error bars. All other parameters are held at their mean and height (a) and Shannon's H (b) are shown in two panels.

FIGURE 6

(a) The mean reproductive tiller count per tussock is shown as a function of fenced and unfenced plots across the study years. (b) The effect of the treatments is displayed on the same scale as panel (a). The error bars for each panel represent one standard error.

FIGURE 7

(a) The mean Shannon's H is shown as a function of fenced and unfenced plots across the study years. (b) The effect of the treatments is displayed on the same scale as panel (a). The error bars for each panel represent one standard error.

FIGURE 8

The number of Orseolia sp. galls, and Tetramesa sp. eclosure holes are shown on panels (a) and (b) y axes, respectively. The y‐axes have the same scale. Insecticide and control bars are grouped by fenced and unfenced treatments on the x‐axes. The error bar is one standard error.

FIGURE 9

For each year, the percentage of plots with infected tussocks is shown. The total samples of tussocks with reproductive tillers evaluated are given above the bar for each year. 2021 had very few reproductive tillers produced. The error bars are one standard error.