#01 2023

Does bush encroachment alter the water cycle? First results from in-situ measurement of evapotranspiration at Mthimkhulu long-term research site

By Tony Swemmer & Tiffany Aldworth, Ndlovu Node, NRF-SAEON

The effect of large-scale changes in vegetation on the freshwater resources of South Africa has long been an area of scientific focus for SAEON. While research on the topic has a long history in the higher rainfall parts of the country, much less has been done in the drier parts. 

The vast semi-arid savannas of southern Africa have experienced a marked increase in the abundance of small trees and shrubs over the past century, and while there are many studies on the impact of this on grass production and rangeland productivity, little is known about how it has impacted the water cycle.

Increasing tree and shrub cover has the potential to intercept more rainfall, reducing the amount of water that infiltrates into the soil, as well as increase transpiration (the process in which much soil water leaves the soil and returns to the atmosphere, via plant leaves). If this occurs over large areas, it has the potential to reduce the flow of local rivers, reduce groundwater recharge and even alter local rainfall patterns.

Figure 1. A simplified depiction of how the cover of trees and shrubs may affect the movement of water through a savanna ecosystem

To determine whether such changes are occurring, the SAEON Ndlovu Node established a long-term ‘bush-clearing experiment’ in the Mthimkhulu Game Reserve in 2015.

Experimental plots of either 60m x 60m or 120m x 120m were cleared of all mopane trees shorter than 4 m, creating ‘treatments’ of very open savanna directly adjacent to control plots of very dense savanna. Soil water sensors were installed in these plots to determine if more rain was getting into, and staying in, the soil in the open savanna plots.

Surface renewal 

PhD student Tiffany Aldworth then set up instrumentation to measure the water losses from  the soil, either directly (through evaporation) or through plant uptake (transpiration).

Surface renewal, a relatively novel method recently trialled by SAEON colleagues at the Grassland-Forest-Wetlands Node, was set up in one treatment plot and its neighbouring control plot. The more widely used eddy covariance method was then utilised for short campaigns to ‘calibrate’ the surface renewal data, ultimately producing nearly three years of continuous measurement of evapotranspiration in two adjacent vegetation types – the first such dataset for savanna ecosystems in South Africa, and probably all of Africa.

Rion Lerm, the technician for the Ndlovu Node, installing Campbell Scientific CS655 soil water sensors at a plot in Mthimkhulu Game Reserve, with help from an assistant from the nearby Mthimkhulu community

PhD student Tiffany Aldworth testing a newly installed surface renewal system for measuring evapotranspiration, with assistance from Ndlovu Node staff Peace Nkuna (left) and Rion Lerm (centre)

While Tiffany’s results show that the amount of water that is lost from the soil is not greatly affected by bush clearing (and by inference, by bush encroachment), some differences in the amount of water stored in the soil were evident. The cover of trees and shrubs may still therefore have an effect on the hydrological cycle by limiting the amount of rainfall that gets into the soil. That is the topic of Tiffany’s next paper – watch this space for those results.

Figure 2. Daily values of evapotranspiration (ET) over the course of three years at Mthimkhulu. The black line shows the ET measured in the control plot with high cover of mopane shrubs, and the grey line in the open, cleared plot. Reproduced from Fig. 8 in Aldworth et al. (2023). The Effect of Woody Encroachment on Evapotranspiration in a Semi-Arid Savanna. Hydrology 2023, 10, 9. https://doi.org/10.3390/hydrology10010009