Now you see me, now you don't: Land use and rainfall pulses interact to determine patterns in small mammal diversity in the Kalahari, South Africa
In arid landscapes, primary productivity is prompted by episodic rainfall. The biological productivity response, in arid systems, to a rainfall event is known as a pulse event that results in an instantaneous change in ecological parameters. These pulse events are high-intensity, short-lived events that result in an abundance of resources. Of the mammals, small mammals are often the quickest to respond to these pulse events numerically and behaviourally. Small mammals are vital to overall ecosystem health and have long been recognised as indicators of environmental change. To test the effect of pulse events on small mammal abundance and diversity, we used the Kalahari Desert as a model system. We compared small mammal data from two surveys during a prolonged drought period (March 2019) and after an extremely wet period (March 2021), at Erin Game Farm and Miershooppan Livestock Farm in the Northern Cape, South Africa. We used a combination of camera trap surveys, walking transects, and live trapping transects to record rodent and meso-carnivore diversity and abundance. Both mammal abundance and diversity were relatively equal between land uses during the drought period survey. However, although there was an observable increase in species diversity and abundance in the wet period, for both land uses, the increase in rodent abundance and diversity in the live trap lines, on the livestock farm were much more prominent. This unexpected result could be attributed to differences in land-use history of the two sites, but further investigation is needed to test this hypothesis. The results of this study provide an understanding of how land use may interact with the effect of the irruption of resources, and subsequent rodent and meso-carnivore response, thus providing vital information for the management of arid landscapes. The first small mammal survey occurred during a period of drought in 2019, from 08 March 2019 to 19 March 2019. The second survey occurred during a period of high rainfall, starting on 25 March 2021, and ended 03 April 2021. Live traps were placed in lines consisting of ten 7.6 cm x 8.9 cm x 22.9 cm Sherman Folding Aluminium traps. The traps were placed approximately 10 m apart from one another in a South to North direction; thereby making 100 m line transects. In addition, on each end of these transects, we placed one 75 cm x 30 cm x 30 cm Standard Humane Cage (SHC) Trap on either end of the transect for the trapping of meso-carnivores. In addition to the traps, the small mammal survey included the daily walking of morning transects with local ‡Khomani San trackers. The observers and trackers were dropped at a different point each day, and walked 1.25 km up from the starting coordinate, 500 m to the right and 1.25 km down thereby resulting in 3 km long transects. Tracks, scat/dung, and visual sightings were observed and recorded. We also performed a camera trap survey with camera traps following the SnapShot Safari protocol.
Steps to reproduce
All analyses were performed in R studio 2022.7.2.576 running R version 4.2.1. To determine species diversity, we calculated separate species accumulation curves for each of the different field methods, using the “BiodiversityR” and “vegan” packages. To circumvent the mismatch in survey effort, we calculated the trap rate rather than total captures. Due to the high number of Sherman traps used, we used the total captures per line, to calculate the trap rate per day. For the SHC traps, we calculated the total captures, per SHC trap, per day. The repeated measures ANOVA with aligned ranks transformation (ART-ANOVA) was performed for both the Sherman and SHC live trap line results to compare the effect of a drought and high rainfall on species abundance (trap rates), between two different land-uses. We used a non-parametric test to avoid issues with non-normal distributions. We compared the capture rates between the two periods to determine if there was a difference between the dry and wet period for the live trap lines and the camera traps. In addition, we compared the capture rates between the sites to determine if the difference in land-use contributed to any observed differences. As the species abundance results came from unrelated samples, i.e drop-off points differed between 2019 and 2021, a Kruskal-Wallis test was performed to determine whether the populations trap rates are identical. On further analyses, to determine which independent variables differ from each other, a Dunn test was performed. For the walking transects, we divided the total observations by 30 to get a track rate per 100 m, per day. The recordings included the spoor count, scat/dung count and visual sightings of all small mammals. The ART-ANOVA was performed for camera trap results to compare the effect of a drought and high rainfall on species abundance (trap rates), between two different land-uses. For the camera traps, we defined capture rate as a capture rate per camera, per site, per survey. We used a 107-day period, thus a total survey season (between two successive camera services), from 20 February to 06 June, for both land-uses in 2019 and 2021. Therefore, the total number of small mammals captured over the 107 days defined the global capture rate, per camera (over 18 cameras in total).
National Research Foundation and the Foundational Biodiversity Information Programme
National Research Foundation