Co-evolution of soils and landscapes in the Anthropocene from natural to intensively managed landscapes

Résumé

Soils provide numerous functions to support natural and human life. Soils and their functions develop over long timescales (decennia to millennia) under influence of environmental properties and drivers such as water flow, vegetation type and topography of the landscape. At the same time, these environmental properties develop too, often under influence of soil properties and processes. This interactive co-evolution of soils and the landscape forms a complex system that can aggravate, or diminish, rates and direction of soil-landscape evolution. In the Anthropocene, a proposed geological Epoch where humans are the main forcing actors, soil-landscape evolution changed substantially under influence of anthropogenic processes, such as deforestation and tillage. In current intensively managed agricultural landscapes in undulating settings, rates of anthropogenic erosion far exceed rates of natural soil development, leading to severe soil and land degradation. Sustainable nature-based land management is crucial to counteract this degradation, and to preserve and restore soil functions for the environment and future generations. The aim of my thesis is to identify and quantify how soils and landscape have evolved and possibly co-evolved during the transition from natural land cover to intensive land management in the Anthropocene.

The first part of this thesis (Chapter 2-3) aims at reconstructing the impact and rates of anthropogenic landscape change on complex agricultural fields. As study site I use the landscape laboratory CarboZALF‑D. CarboZALF‑D is a kettle-hole catchment of 4 ha with elevation differences up to 8 meters, located in north-eastern Germany. The catchment is characterized by complex small-scale topography, heterogeneities in the hydrological system and a long history of agricultural use. The colluvium in the closed kettle hole catchment provides a complete geo-archive of landscape change. In Chapter 2 we reconstruct the paleosurface of study site Carbo-ZALF-D prior to the anthropogenic erosion. We used an extensive dataset of soil descriptions, which enabled a detailed spatial estimate of erosion and deposition by estimating erosion based on soil profile truncations and deposition based on colluvium thickness. The paleosurface shows a high variation in topographic properties and suggests that natural soils and landscapes contain considerable spatial heterogeneity. In Chapter 3 we reconstruct the rates of deposition in Carbo-ZALF-D using Optically Stimulated Luminescence (OSL) dating. We present a novel methodology to apply OSL dating in colluvial sediments, where the soil chronology gets disturbed by reworking by ploughing after deposition. Our results show a 100-fold increase in deposition rates, starting around 5000 years ago. This increase does not solely represent increased erosion in the catchment, but is also caused by indirect effects of agricultural drainage. The kettle hole shows a complex spatiotemporal pattern of colluvial infilling and landscape evolution, which we were only able to reconstruct using a high OSL sampling density and extensive soil geomorphic research. 

The second part of this thesis aims at simulating the evolution of soils and landscapes under varying climatic and anthropogenic forcing. In Chapter 4 we review the role of water as dominant driver in natural soil and landscape evolution and its potential as driver in simulations with soil-landscape evolution models (SLEMs). Water plays a pivotal role in soil and landscape evolution, by transporting and transforming soil material and facilitating vegetation growth. In turn, surface and subsurface flow paths of water are controlled by soil and landscape properties. The co-evolution of soils, topography and the hydrological system is essential for understanding the response of soils and landscapes to changes in climate. However, this co-evolution can currently not be simulated over long timescales with SLEMs due to several conceptual and methodological challenges. We provide partial solutions for these challenges. In Chapter 5 we utilize these partial solutions to develop our SLEM HydroLorica. HydroLorica simulates soil and landscape evolution with various dynamic drivers such as water flow, vegetation type and land use. We included additional essential processes such as tree throw, soil creep and tillage. We use HydroLorica to simulate the evolution of soils and landscape under various rainfall and land-use scenarios for an artificial undulating landscape. The results show that in natural systems, rainfall amount is the dominant factor controlling soil and landscape heterogeneity, while for agricultural systems landform explains most of the variation. The cultivation of natural landscapes increases soil heterogeneity, but also increases correlations between soil and terrain properties. Our results confirm that humans have become the dominant soil forming factor in intensively managed landscapes. 

In the third part of this thesis (Chapter 6), I synthesize the findings from the research chapters to meet the objectives of this thesis. I critically evaluate the developed reconstruction methods in Chapters 2 and 3 and compare them with other potential methods. The development of HydroLorica in Chapters 4 and 5, with water flow as explicit driver and with increased process coverage, is a big step forward in soil-landscape evolution modelling. A combination of reconstruction and simulation methods is essential for developing and testing hypotheses of soil-landscape co-evolution. Soil-landscape evolution in natural and intensively managed landscapes have different characteristics due to different driving forces and dominant processes. In natural landscapes, soils develop to patterns where individual soils might be disturbed occasionally, but where the average properties are stable. In intensively managed landscapes, disturbance rates are much higher than in natural settings. As a consequence, slowly developing soil properties degrade, while fast-developing soil properties can form a new equilibrium. The co-evolution of soils and landscapes that occurs in natural settings is often controlled by biotic processes. In agricultural settings, humans control vegetation type and aggravate erosion processes through tillage. As a consequence, co-evolution does not occur in the sense that it does in natural settings, because interactions between landscape components are missing. However, the management of soils and landscapes is often adapted to counteract unintended changes to soils and landscapes under earlier management. In intensively managed landscapes, land management may thus co-evolve with the rest of the landscape.