Dust emission from croplands in the Free State, South Africa

The global dust load showed a large increase during the last century due to climate change and the expansion of vulnerable land, both of which are caused by human modifications. The increase in vulnerable land, both in size as in intensity, is mainly attributed to the increase in agricultural areas and agricultural intensification. Dust emission has both an onsite effect due to the degradation of the emitting area, and an offsite effect on human health and climate. The degradation of land is especially relevant for agricultural areas where crop yield can strongly diminish due to the depletion of clay, silt, and nutrients from soils.
The semi-arid Free State province has been identified as the largest emitter of dust in South Africa, which is caused by the large-scale agriculture, climate, and soil type. Dust storms have the potential to reach the densely populated Gauteng province causing negative consequences on human health and well-being. Dust events in the Free State show a strong seasonality that is attributed to the agricultural practices that leave soils bare and vulnerable to erosion after harvesting. However, the large differences in dust events per year indicate that additional surface characteristics control the emissions from these harvested croplands. One of the primary potential controls that farmers have on the emissivity of the land is the management of soil crusts, but, the role of soil crusts on sandy soils is often not considered. Therefore, this thesis will address the possible role of soil crusts on the dust emission from the Free State croplands.
This thesis examines the formation of crusts by rainfall, the dust emissions from cropland soils, and the surface characteristics that control this erosion. The main instrument used to measure the erodibility of a surface is the Portable In-Situ Wind Erosion Laboratory (PI-SWERL). The comparability of this portable instrument was assessed by a cross-comparison with a traditional straight-line wind tunnel. The threshold friction velocity of sandy surfaces was similar for the two instruments, whereas the threshold friction velocity of loamy sand indicated that the PI-SWERL is a more precise instrument that is capable of detecting the initial, small PM10 emissions from a surface.
To determine the potential for crusts to form on the sandy cropland soils, rainfall experiments combined with shear strength measurements were performed on Free State soils. The results showed that significant crusts develop within 15 mm of rainfall, and shear strengths similar to those measured in the laboratory were observed in the field. PI-SWERL measurements showed that these experimental crusts can limit the PM10 emission flux from 10.53 and 3.87 mg m-2 s-1 Luvisol and Arenosol soils, respectively, to below 0.03 mg m-2 s-1 for both. The addition of abraders increased the emission from a crust to 0.43 and 0.31 mg m-2 s-1 for Luvisol and Arenosol, respectively.
The strong effect of crusts on emissions have been compared to field measurements, which showed a similar potential to minimize dust emissions, but also are complex interaction on the surface that defines this influence. The average emission of crusted surfaces was 0.476 mg m-2 s-1 (standard II deviation = 0.348, min = 0.004, max = 1.401 mg m-2 s-1) at a friction velocity of 0.59 m s-1, whereby the presence of abraders showed a power relationship to the emission from these surfaces. The emission from loose surfaces ranged between 1.646 mg m-2 s-1 (standard deviation 0.980, min = 0.291, max = 5.974 mg m-2 s-1), with a linear relationship between the emission and the clay and silt content.
The initial sensitivity of an agricultural field to wind erosion needs to be considered when assessing the surface conditions under which crusts could play a minimizing role in the emission of dust. This is controlled by the soil cover, such as vegetation or degrading crop straw and stubble that is left after harvesting. Four fields, with a range of soil cover and crust characteristics, have been measured in the field. The soil cover has been quantified using Unmanned Aerial Vehicle (UAV) image analyses. The erodibility of the soils was characterised by the horizontal sediment flux and the saltation threshold. The soil cover differed from 11% for a harvested groundnut to 66% for a harvested maize field, the latter being the most common crop type in the Free State. This data shows the high initial importance of soil cover on the wind erosion from a field, whereby the sediment flux from the maize field was 11 and 187 times lower than that of fallow and groundnut fields. Considering the relatively high sediment flux from the fallow and low soil cover fields, crust and abrasion management should be considered on such surfaces.
Some surfaces showed a depletion of clay and silt, which is evidence of land degradation caused by dust emission. This depletion could eventually lead to a lower yield and the need for more fertilizers. Furthermore, an enrichment in certain allergens and pathogens has been found in the suspended dust from Free State croplands. This shows the relevance of minimizing dust, both for the emitting region and the offsite areas that dust eventually reaches.
Future studies should investigate the relationship between the sediment flux and the PM10 flux since this relationship is not known and could differ per field and soil type. Furthermore, the influence of roughness needs to be assessed because roughness is generally known to decrease the emission from a surface. However, to create roughness, it is required to disturb existing crusts, making it uncertain if such activities could increase or decrease overall emissions. Lastly, the implications raised by this thesis are not only relevant for Free State but can also be considered for other cropland areas with strong seasonality in cover and moisture.