Multi-scale dynamics of mixed sand and gravel deltaic coasts

Abstract

Deltaic systems are located at the transition between fluvial and maritime environments. They all have high environmental, economic and social importance, and respond rapidly to both natural and human-driven changes. On the other hand, mixed sand and gravel (MSG) beaches are common in previously para-glaciated coastal regions and coasts with steep hinterlands, and are widespread in the UK, Denmark, Canada, New Zealand and Mediterranean countries. They are also found when nourishment projects use gravel to protect eroded sandy beaches. Despite their societal importance, the research advances on gravel and MSG beaches have been limited compared to those on sandy beaches. The main objective of this thesis is to analyse, characterize and model the dynamics of MSG deltaic coasts based on a multi-scale investigation carried out in the Guadalfeo delta, a Mediterranean delta in a semi-arid and high-mountain basin of southern Spain. To meet this overall objective, five specific goals were proposed, which were addressed in five different chapters. First, the effects of the construction of a reservoir 19 km from the mouth on the dynamics of the delta are analysed in Chapter 3. The sediment volume transported as bedload and accumulated in the delta was estimated under two scenarios by means of a calibrated hydrological model: a managed scenario, considering the flows drained by the dam, and an unmanaged scenario, considering the absence of such infrastructure. Bathymetric and topographic measurements were analysed and correlated with the fluvial and maritime forcing agents. Results indicated that the reservoir has significantly modified the dynamics downstream: the coast has lost almost 0.3 hm3 of sediments since the entry into operation of the dam, generating a 1.4-km coastline retreat around the mouth, with a maximum retreat of 87 m (92% of the initial). Under unmanaged conditions, more than 2 hm3 of bedload would have reached the coast. Based on the results, three new management scenarios of flows drained by the dam, in combination with bypassed sediment from the reservoir, were proposed to prevent more severe consequences in the delta and the silting of the reservoir. The proposed methodology for new management scenarios can be extended to other worldwide deltas and represents an advanced tool for decision making. Secondly, Chapter 4 details the spatial and temporal variability of the river mouth, focusing on the influence of submerged morphological changes on wave propagation and longshore sediment transport (LST). Bathymetric measurements were carried out over a 15-year period (1999 - 2014), a wave propagation model (Delft3D-Wave) was calibrated and applied, and the complete littoral drift time series was obtained using statistical downscaling techniques. Results showed that the river damming led to coastline retreat and bed-level erosion up to 3 m along a 1-km section around the river mouth, with maximum erosion rates in excess of 760 m3/m. These subtidal morphological changes reduced wave refraction and led to higher breaking wave energy. Variations in wave climate during the study period also played a role in influencing the coastline dynamics. Although the erosion around the river mouth has decreased since 2008, partly due to a sediment pulse in 2010, eastward LST rates under westerly storm wave conditions have significantly increased since then. This has led to the propagation of the sediment deficit towards the east of the mouth, endangering urban developments at this location. This chapter provides insights into the shift from wave-river dominated deltas towards deltaic coasts increasingly controlled by wave directionality and LST, and represents an advance in the understanding of the dynamics of many worldwide deltas where the river sediment supply has decreased due to human activities. The changes in the morphology and sedimentology of the MSG beach forced by wave and water-level variations as well as human intervention through nourishment are studied in Chapter 5. Monthly and storm event-driven field surveys, consisting of topographical measurements and sediment sampling, were carried out over a one-year period (October 2013 - September 2014). Three prevailing sediment fractions (sand, fine gravel and coarse gravel) and two end-member morphological states of the upper beach profile (convex with multiple berms and concave with a single storm berm) were identified. Between them, several transitional profiles were formed, characterized by developing berms that progressively overlapped, generating sediment variability both across the beach profile and with depth. Results indicated that the total run-up (including water-level) reached during an event represents a more accurate threshold than wave height for differentiating between erosional and depositional conditions. They also suggested that MSG coasts recover faster from storm erosion than sandy beaches. The long-term benefit of the artificial nourishment was limited and this was attributed to the too fine nourished sediment used and its placement too high on the beach profile. This chapter deepens into the knowledge of the morpho-sedimentary dynamics of MSG beaches under varying wave and water-level conditions. Chapter 6 addresses the evolution of the coast forced by an artificial nourishment project ended in December 2014 through the analysis of field observations and the modelling of hydro- and morphodynamics. The beach profile and coastline were periodically surveyed over a six-month period; the Delft3D-Wave model was applied; four LST equations were tested through comparisons with bathymetric data; and the one-line model was applied between topographic surveys. Results indicated that severe coastline retreat (dry beach area loss >208 m2/day) occurred during the 45 days following the intervention. This was mainly attributable to the morphology of the nourished coastline, the different characteristics of the nourished sediment compared to natural sizes, and the occurrence of an intense westerly storm. The Van Rijn formulation provided the best fits to the observed volumetric changes, obtaining modelled/measured ratios of 93.1% and 77.4% for the two study beach profiles. The outputs of the one-line model based on the Van Rijn approach were also the best, with RMSEs decreasing during the study period and lower than 4.6 m over the last 3 months. These results show that the applied methodology constitutes a management tool for modelling the evolution of MSG coasts. Finally, the beach profile response forced by storm waves from varying directions is investigated in Chapter 7. Beach morphology was monitored over a 36-day period between January and February 2015, and profile response was compared to model predictions using the XBeach-G model and a LST formulation. XBeach-G was applied over 2-day periods of low energy, south-westerly (SW) storm and south-easterly (SE) storm conditions, and was coupled to LST using a parametric approach which distributes the LST across the swash, surf and nearshore zones. The Delft3D model was used to obtain the inshore conditions required to drive XBeach-G and the LST formulation. The storm response was clearly influenced by the free-board (difference between the height of the berm and the total run-up) and was strongly dependent on storm-wave direction, with the SW and SE storms eroding and building up the surveyed area, respectively. Model results indicated that XBeach-G on its own is capable of reproducing the beach response under SW storm conditions (BSS>0.95), but not under SE storms due to the higher LST gradients. The combination of XBeach-G and LST improved the fits to measured profiles under both SW (BSS>0.96) and SE (BSS>0.88) storms. This coupled approach represents an extension of XBeach-G to make it more suitable for coasts highly influenced by both cross-shore and LST, such as beaches with different coastline orientations and/or forced by varying wave directions.