Impact of bioturbation on sediment redistribution in coastal Chile - As estimated by combining remote sensing, machine learning and semi-empirical modelling
The burial activity of terrestrial bioturbators influences the microtopography, surface roughness, and physical properties of the soil. By reworking sediments, bioturbators increase soil permeability and porosity, which has implications for infiltration and erosion rates. The construction of undergr...
|PDF Full Text
No Tags, Be the first to tag this record!
|The burial activity of terrestrial bioturbators influences the microtopography, surface roughness, and physical properties of the soil. By reworking sediments, bioturbators increase soil permeability and porosity, which has implications for infiltration and erosion rates. The construction of underground tunnels distributes and concentrates nutrients and has a particularly positive effect on carbon storage in the soil. Previous studies have left several research gaps. The studies focused only on the habitat preferences of individual species and did not consider the varying amount of excavated sediment and the building density of individual species. It remains unclear which environmental parameters within the catchment area are primarily associated with the high density and distribution of all existing bioturbator structures. Furthermore, the previous authors did not address the daily sediment excavation dynamics by the animal, whether and how it is related to sediment redistribution driven by catchment-wide precipitation, and how much sediment the bioturbators transport to the surface throughout the year.
My dissertation was part of the EarthShape consortium with the overarching research question of how microorganisms, animals, and plants influence the shape and development of the Earth's surface. The study was conducted at four study sites along the Chilean coastal cordillera: arid Pan de Azúcar, semi-arid Santa Gracia, Mediterranean La Campana, and humid Nahuelbuta. The workflow consisted of three work packages with the ultimate goal of determining the catchment-wide effects of bioturbation.
Within the first work package, I tested whether the density of burrows and the distribution of structures can be predicted by vegetation patterns calculated from UAV and WorldView-2 data. Then I used the best model for catchment-wide prediction. Within the second work package, I tested whether bioturbator-driven sediment redistribution depends on precipitation-driven sediment redistribution. For this purpose, I deployed several time-of-flight-based cameras to monitor sediment redistribution on the trench surface and around the trench. In the third work package, I integrated bioturbation into a soil erosion model. Then I determined the influence of bioturbators on sediment distribution and the environmental parameters that determine the extent of this influence.
My results showed that the distribution of structures created by bioturbating animals depends on vegetation patterns. The density of burrows created by bioturbating animals was best predicted by in-situ measured vegetation cover as well as the diameter and height of shrubs. In the arid and semi-arid zones, cactus height and cover were important predictors, while in the humid zone, tree trunk diameter and cover were selected by the model. However, plant species diversity was important in all climatic zones. When predicting burrow density using UAV images, indices of vegetation heterogeneity were also important. The density of structures increased with shrub, herb, and cactus cover in all climatic zones and decreased with tree canopy cover in the humid climate zone. The density of invertebrate structures was higher in rockier areas with less vegetation at all sites. Finally, a vegetation index describing high leaf area index was an important predictor. The distribution of structures throughout the catchment area was best predicted by the WorldView-2 NIR band and NDVI, as well as individual vegetation land cover classes. Topographic features derived from LiDAR data were not selected as important predictors, except for aspect.
Secondly, the results showed that sediment redistribution triggered by bioturbators depends on rainfall-triggered redistribution. Immediately after rainfall events in the Mediterranean climate zone, increased sediment export by the animals was observed: the animals were observed reconstructing their structures after the rains and simultaneously excavating more additional sediment to the surface. In contrast, in the arid climate zone, sediment export was mostly not preceded by rainfall events.
The results confirmed that the environment determines the extent of the impact of bioturbation on rainfall-induced sediment redistribution throughout the catchment area. The results showed that the key environmental parameters were elevation, surface roughness, slope, and vegetation cover derived from NDVI. Bioturbation increased sediment erosion in areas where erosion processes dominate (steep slopes, strong gradients, low surface roughness, low vegetation cover), and similarly increased sediment accumulation in areas with natural sediment deposition (high surface roughness, high vegetation cover, low slope).
The model output demonstrated that bioturbation intensifies sediment erosion. Bioturbation amplified sediment erosion in all climatic zones except the humid zone. Monitoring the structures showed an increase in erosion of over 300% compared to the areas where the structures were embedded. According to the results of the soil erosion model, bioturbation had the strongest impact on erosion in the Mediterranean zone, followed by the arid and semi-arid zones. The effects of bioturbation were not significant in the humid zone.
To assess long-term impacts, bioturbation needs to be integrated into landscape development models. However, these models have assumed a uniform distribution and spatial and temporal effects of bioturbation. My results demonstrated that the effects of bioturbation on sediment redistribution are not temporally and spatially consistent, and the distribution of bioturbation is not uniformly associated with vegetation. To realistically predict the long-term effects of bioturbation, the estimated spatial and temporal variations from this study need to be considered.