Getting a grip on hydrological and sediment connectivity | Land Portal

Resource information

Date of publication: 
December 2017
Resource Language: 
ISBN / Resource ID: 
NARCIS:wur:oai:library.wur.nl:wurpubs/525833
Copyright details: 
Open Access, this refers to access without restrictions, and without financial incentives. Access to the resource is gained directly, without any obstacles. From info:eu-repo/semantics/openAccess

Land degradation is a large problem worldwide, especially in agricultural areas. Between 1-6 billion ha of land worldwide is affected by land degradation. With an increasing world population, more food production is needed and, therefore, more land is converted into agricultural areas. This conversion of land to agricultural areas, in turn, leads to more land degradation. Some common forms of land degradation are desertification, salinization and soil erosion by water. The negative effects of soil erosion have been recognized for a long time. Since the early 20th century, researchers have tried to quantify soil displaced due to water, and to measure and model the efficiency of management strategies.

The implications of problems with upscaling, wrong process representation and equifinality include the difficulty to properly predict sediment sources, pathways and sinks within catchments. These problems then can translate into the implementation of sub-optimal management strategies. To deal with these non-linear processes and the lack of proper representation of water and sediment sources, pathways and sinks, the concept of connectivity was developed. Currently, many definitions of connectivity have been proposed, although the definition most used is that of hydrological connectivity by Pringle (2003): ‘Hydrologic connectivity is the water-mediated transport of matter, energy and organisms within or between elements of the hydrologic cycle’.

A unified theory on what constitutes connectivity and how connectivity should be measured or inferred remains one of the biggest challenges within catchment science. In addition, it is unclear whether connectivity should be an output or an input of a model and if an input, whether this should be added explicitly or implicitly. The main objective of this thesis was, therefore, to assess and quantify hydrological and sediment connectivity in a meaningful way, which can further our understanding of hydrological and sediment transport processes and catchment system dynamics.

The study was carried out in three catchments in Navarre, northern Spain. Two catchments, ‘Latxaga’ and ‘La Tejeria’, are agricultural catchments with sizes of 2.07 km2 and 1.69 km2, respectively. The ‘Oskotz Forestal’ catchment is a (semi-)natural catchment, with a size of 5.05 km2. Land cover in the agricultural catchments is mainly winter wheat and barley, while in the Oskotz catchment it is grassland and forest. Latxaga and La Tejeria are mainly underlain by marls and within La Tejeria some sandstone is also present. The geology in Oskotz is characterised by an alternation of marls and sandy limestone.

In chapter 2, I used networks (graph theory) to characterise and quantify overland flow connectivity dynamics on hillslopes in a humid sub-Mediterranean environment by using a combination of high-resolution digital-terrain models, overland flow sensors and a network approach. Results showed that there are significant differences between overland flow connectivity on agricultural areas and semi-natural shrubs areas. Significant positive correlations between connectivity and precipitation characteristics were found. Significant negative correlations between connectivity and soil moisture were found, most likely due to soil water repellency and/or soil surface crusting. The combination of structural networks and dynamic networks for determining potential connectivity and actual connectivity proved a powerful tool for analysing overland flow connectivity.

In chapter 3, I determined the functioning of hillslope-channel connectivity and the continuation of transport of these sediments in the channel. To determine this functioning, I obtained data on sediment transport from the hillslopes to the channels while simultaneously looking at factors that influence sediment export out of the catchment. For measuring hillslope-channel sediment connectivity, Rare-Earth Oxide (REO) tracers were applied to a hillslope in the Latxaga catchment preceding the winter of 2014-2015. The results showed that during the winter there have been no sediments transported from the hillslope into the channel. Analysis of precipitation data showed that although total precipitation quantities did not differ much from the mean, the precipitation intensities were low. Using a Random Forest (RF) machine learning method, I showed that hillslope-channel connectivity in Latxaga is dominated by sediment mobilisation during large (high intensity) precipitation events. Sediments are for a large part exported during those events. Large events also leave behind large amounts of sediments in and near the channel, which is gradually removed by small events.

In chapter 4 I demonstrated that existing data can be used to assess governing factors of connectivity, and how these factors change over time. Data from three catchments in Navarre, Northern Spain, were used to assess factors that influence hydrologic and sediment connectivity. These factors were used as components in a spatially-lumped linear model for discharge and suspended-sediment yield. Three components of connectivity were distinguished: topographical, biological and soil. Changes in the topographical component for the studied periods were considered relatively small, and, therefore, kept constant. Changes in the biological component were determined using the Normalised Difference Vegetation Index. Changes in the soil component were assessed using an Antecedent Precipitation Index. Nash-Sutcliffe model efficiency coefficients were between 0.49 through 0.62 for the discharge models and between 0.23 through 0.3 for the sediment-yield models. I recommended applying the model at smaller spatial scales than catchment scale to minimize the lumping of spatial variability in the components.

In chapter 5, the objective was to better understand the implications of model calibration at different spatial scales on the simulation of hydrology and sediment dynamics of an agricultural catchment. I applied the LAPSUS-D model to the Latxaga catchment. The model was calibrated and validated (4 years: 2011-2015) using three datasets at varying spatial scales: hillslope, catchment and the combined dataset (combined-calibrated model). The hillslope-calibrated model showed mainly infiltration-excess overland flow, the catchment-calibrated mainly saturation-excess overland flow at the footslopes and the combined-calibrated model showed saturation-excess overland flow from the midslopes to the footslopes. For hydrology, the combined-calibrated model simulated the large discharge peaks best, while at the hillslope scale, the hillslope-calibrated model performed best. The hillslope-calibrated model produced the highest model efficiencies for sediments, for calibration (0.618) and validation (0.269). The hillslope-calibrated model was the only model that showed observed gully erosion on a high-resolution DEM and displayed channel sediment dynamics. However, absolute quantities of erosion and deposition within the catchment were too high. The results show that modellers need to be aware of problems associated with automatic calibration, over-calibration and not incorporating measured data at multiple spatial scales. We advocate incorporating runoff and sediment tracing data at multiple scales whenever this is possible and to, furthermore, carry out specific measuring campaigns towards this end, ultimately to get a more comprehensive view on hydrological and sediment connectivity within a catchment.

The combination of chapters in this thesis showed that the connectivity concept is useful for a wide range of studies, from hillslope scale to catchment scale. Using the concept, I was able to determine sediment dynamics for a humid-Mediterranean catchment and show that this behaviour is different than previously thought.

Depending of the aim of the study, various concepts of connectivity are useful. Different geologic and climatic settings cause large differences in catchment (sediment) dynamics. It might, therefore, not be necessary, or even possible, to strive for a single, unifying conceptual framework for connectivity. Instead, a collection of frameworks for different settings should be developed. These frameworks should, however, always aim at helping to understand which measurements need to be taken and which type of models and indices should be used for that particular setting.

It is my honest opinion that connectivity is definitely a useful concept to advance our knowledge on water and sediment transport processes further. However, careful consideration is also required as this particular concept will not necessary provide the ultimate explanation and insights in dynamic behaviour within watersheds around the world. The gap between the different spatial and temporal scales is too complex to be bridged with a single concept like connectivity. However, the many studies about connectivity that will be published in the near future will be able to advance knowledge on water and sediment transport processes.

Authors and Publishers

Author(s), editor(s), contributor(s): 

Masselink, Rens J.H.
Coen Ritsema
Sjoerd van der Zee
Saskia Keesstra
Arnaud Temme

Publisher(s): 
Wageningen University logo

Wageningen University & Research is a collaboration between Wageningen University and the Wageningen Research foundation. 

That is the mission of Wageningen University & Research. A staff of 6,500 and 10,000 students from over 100 countries work everywhere around the world in the domain of healthy food and living environment for governments and the business community-at-large.

Data provider

National Academic Research and Collaborations Information System (NARCIS) is the main Dutch national portal for those looking for information about researchers and their work. NARCIS aggregates data from around 30 institutional repositories. Besides researchers, NARCIS is also used by students, journalists and people working in educational and government institutions as well as the business sector.

 

Share this page