Final Report Summary - BRAIDSIDEEARTH (The Braided Side of the Earth: modelling the long-term morphological impact of dams on the gravel-bed braided rivers of New Zealand to support restoration of the heavily-impacted European rivers.)
The BraidSideEarth project aims to model the long-term impact of dam construction and operation over braided morphologies, by the development and application of suitable numerical modelling approaches.
Braided rivers, once common in Europe in the Alpine and pre-Alpine region, where the conditions for their development (slopes yielding high sediment loads and determining high-energy flows) existed, in the last century have generally experienced a transition towards single-thread morphologies, together with narrowing, incision and vegetation spread. Such a change has been the consequence of multiple anthropogenic stressors applied to rivers. Among these, the construction and operation of dams (for hydropower, irrigation and flood safety) has played a key role, since dams are able to alter two main controls over the river shape, namely the flow regime and sediment yield, and, through the altered flow regime, to favour vegetation spread. Loss of numerous braided reaches has implied a severe loss in terms of the landscape, morphological and ecological values that braided morphologies could fulfill.
Despite the relevance of the topic, there is no modelling tool able to forecast the long-term evolution trajectory of a river morphology in response to damming. Hence, the construction and application of such a tool is the main goal of the present project.
In other regions of the Earth such as New Zealand’s South Island the impact of anthropogenic activities on braided morphologies has been comparatively lower. This makes New Zealand the ideal setting to study braided rivers in quasi-pristine conditions (from European perspectives), to observe the natural processes of braiding and to develop and train numerical models suitable for physics-based prediction. Therefore, the project has benefited of a two-year outgoing phase in New Zealand, before a one-year reintegration phase at the University of Trento (Italy).
The project’s modelling objective defines four research tasks (RTs), which are hereby summarised.
- RT1: construction of a database of real-world gravel-bed braided rivers which have undergone morphological change due to damming, and identification of the metrics suitable to quantify these changes;
- RT2: development of a suitable physics-based numerical model;
- RT3: application of the model to quantify the impact of dams on braided morphologies;
- RT4: quantification of river response to various hypothetical managing scenarios to devise restoration strategies and mitigation strategies.
The work performed and the results achieved with respect to each RT are as follows.
RT1- A database of gravel-bed braided rivers which have undergone morphological change has been collected and analysed. Metrics suitable to quantify such change have been identified, namely: bed degradation, change in braiding index, floodplain narrowing, and vegetation spread. These four metrics have been used to characterise the morphological change of three selected river systems (Waitaki, New Zeland; Piave, Italy; Dunajec, Poland). Results show that all these rivers have experienced similar changes; however the trajectory of each river was determined by a different blend of anthropogenic impacts applied, which impacted each river’s morphological process in a different way: in other words, rivers experienced similar changes, but for different reasons. This implies that, to restore the natural processes and shapes of one river, the actual history of alterations in that particular river must be carefully reconstructed.
RT2- The pre-existing hydro-morphodynamic code GIAMT2D (accounting for 2D flows and bedload sediment transport) has been enhanced with a bank erosion module and a state-of-the-art vegetation module. The enhanced GIAMT2D model is the only one presently available that accounts for all the processes relevant to studying the impact of damming on braided morphologies, and hence suited for the project.
RT3- The model has been applied and tested in a simplified configuration and in controlled conditions against literature data (the outcomes of the laboratory experiments performed by Tal and Paola (2010)). In that experiment, a braiding river initially created in a bare sand flume bed was progressively reduced to an almost single-thread river due to the channel bank-strengthening action of vegetation (alfa-alfa grass sprouts) which had been introduced into the flume. A calibrated numerical model has proven able to give good quantitative reproduction of the observed dynamics of vegetation spread and morphological change in the laboratory experiment. This represents the first ever numerical reproduction of a physical experiment focused on the mutual interactions between morphological change and change in vegetation cover in rivers.
Following the same principles, the model has been then applied to reproduce spread of invasive vegetation accompanied by reduction of morphological complexity and activity observed in the lower Waitaki River (New Zealand) in the period between 1936 and 1950, due to the introduction of non-native vegetation which spread following the adoption of a dam-dampened flow regime after the closure of Waitaki Dam in 1936 (Hicks et al., 2003).
RT4- Scenario applications have been carried out on the aforementioned Waitaki River in New Zealand, heavily impacted by dams and invasive vegetation to devise restoration strategies. Simulations have focussed on studying the river evolution that would have occurred if Waitaki Dam had not been constructed, or if intermediate management scenarios (yielding intermediate flow regimes between the one “with Waitaki Dam” and the one “without Waitaki Dam”) had been adopted
Results show that the adopted flow regime plays a key role in allowing or limiting vegetation spread and hence the fate or the river morphology; that without construction of Waitaki Dam, the river would have been substantially spared from the morphological transition which has occurred; and that, even after severe alteration due to the introduction of alien vegetation, the adoption of a more or entirely natural flow regime could still provide significant relief to the river morpho-ecological environment.
A summary of the project results is given in the project website (https://sites.google.com/view/braidsideearth/). Current email contact of the Fellow is Gu.Stecca@niwa.co.nz.
The results achieved in different areas (geomorphology literature review, model development, model application) are complementary to each other and help unravel the mechanisms and dynamics of morphological change in impacted braided rivers. The model, trained on laboratory data and/or field data (when available) can be used to predict river behaviour in a countless number of different situations by simply changing inputs and controls. This helps scientists stretch their prediction ability well beyond the single cases used to train models, as our scenario applications show. These results are making in-roads towards the development of geomorphology as a quantitative and predictive discipline, and will potentially help devising mitigation and restoration strategies for impacted river systems.
Natural end users of our work are river managers, professionals, administrations interested in understanding the river evolution up to the present and to assess the impact of planned (changes in) management rules in the future. The numerical model developed within the project, despite being a research tool, represents the prototype for tools which we foresee will be of wider use among the above categories of stakeholders in the medium term (from a few years to a couple of decades). Before such a model can become a standard tool, however, some difficulties must be overcome. These difficulties are related to the high-demanding nature of research model both in terms of data and parameters to inform correct functioning of all the submodels (for bank erosion, vegetation) and to the very high demand of computational power needed to run long-term simulation in domains covering reasonably long and wide river reaches. Both these kinds of difficulties are foreseen to be progressively overcome in time with better spread and testing of these tools (which will help better understand how to correctly inform these with parameters), progressively improved data availability, and ever-increasing available computational power. In the meantime, there is room for a huge number of applied projects aiming to let the proposed model improvements leak into the most known and commonly used modelling frameworks, and to apply these tools to more test and pilot cases.
Braided rivers, once common in Europe in the Alpine and pre-Alpine region, where the conditions for their development (slopes yielding high sediment loads and determining high-energy flows) existed, in the last century have generally experienced a transition towards single-thread morphologies, together with narrowing, incision and vegetation spread. Such a change has been the consequence of multiple anthropogenic stressors applied to rivers. Among these, the construction and operation of dams (for hydropower, irrigation and flood safety) has played a key role, since dams are able to alter two main controls over the river shape, namely the flow regime and sediment yield, and, through the altered flow regime, to favour vegetation spread. Loss of numerous braided reaches has implied a severe loss in terms of the landscape, morphological and ecological values that braided morphologies could fulfill.
Despite the relevance of the topic, there is no modelling tool able to forecast the long-term evolution trajectory of a river morphology in response to damming. Hence, the construction and application of such a tool is the main goal of the present project.
In other regions of the Earth such as New Zealand’s South Island the impact of anthropogenic activities on braided morphologies has been comparatively lower. This makes New Zealand the ideal setting to study braided rivers in quasi-pristine conditions (from European perspectives), to observe the natural processes of braiding and to develop and train numerical models suitable for physics-based prediction. Therefore, the project has benefited of a two-year outgoing phase in New Zealand, before a one-year reintegration phase at the University of Trento (Italy).
The project’s modelling objective defines four research tasks (RTs), which are hereby summarised.
- RT1: construction of a database of real-world gravel-bed braided rivers which have undergone morphological change due to damming, and identification of the metrics suitable to quantify these changes;
- RT2: development of a suitable physics-based numerical model;
- RT3: application of the model to quantify the impact of dams on braided morphologies;
- RT4: quantification of river response to various hypothetical managing scenarios to devise restoration strategies and mitigation strategies.
The work performed and the results achieved with respect to each RT are as follows.
RT1- A database of gravel-bed braided rivers which have undergone morphological change has been collected and analysed. Metrics suitable to quantify such change have been identified, namely: bed degradation, change in braiding index, floodplain narrowing, and vegetation spread. These four metrics have been used to characterise the morphological change of three selected river systems (Waitaki, New Zeland; Piave, Italy; Dunajec, Poland). Results show that all these rivers have experienced similar changes; however the trajectory of each river was determined by a different blend of anthropogenic impacts applied, which impacted each river’s morphological process in a different way: in other words, rivers experienced similar changes, but for different reasons. This implies that, to restore the natural processes and shapes of one river, the actual history of alterations in that particular river must be carefully reconstructed.
RT2- The pre-existing hydro-morphodynamic code GIAMT2D (accounting for 2D flows and bedload sediment transport) has been enhanced with a bank erosion module and a state-of-the-art vegetation module. The enhanced GIAMT2D model is the only one presently available that accounts for all the processes relevant to studying the impact of damming on braided morphologies, and hence suited for the project.
RT3- The model has been applied and tested in a simplified configuration and in controlled conditions against literature data (the outcomes of the laboratory experiments performed by Tal and Paola (2010)). In that experiment, a braiding river initially created in a bare sand flume bed was progressively reduced to an almost single-thread river due to the channel bank-strengthening action of vegetation (alfa-alfa grass sprouts) which had been introduced into the flume. A calibrated numerical model has proven able to give good quantitative reproduction of the observed dynamics of vegetation spread and morphological change in the laboratory experiment. This represents the first ever numerical reproduction of a physical experiment focused on the mutual interactions between morphological change and change in vegetation cover in rivers.
Following the same principles, the model has been then applied to reproduce spread of invasive vegetation accompanied by reduction of morphological complexity and activity observed in the lower Waitaki River (New Zealand) in the period between 1936 and 1950, due to the introduction of non-native vegetation which spread following the adoption of a dam-dampened flow regime after the closure of Waitaki Dam in 1936 (Hicks et al., 2003).
RT4- Scenario applications have been carried out on the aforementioned Waitaki River in New Zealand, heavily impacted by dams and invasive vegetation to devise restoration strategies. Simulations have focussed on studying the river evolution that would have occurred if Waitaki Dam had not been constructed, or if intermediate management scenarios (yielding intermediate flow regimes between the one “with Waitaki Dam” and the one “without Waitaki Dam”) had been adopted
Results show that the adopted flow regime plays a key role in allowing or limiting vegetation spread and hence the fate or the river morphology; that without construction of Waitaki Dam, the river would have been substantially spared from the morphological transition which has occurred; and that, even after severe alteration due to the introduction of alien vegetation, the adoption of a more or entirely natural flow regime could still provide significant relief to the river morpho-ecological environment.
A summary of the project results is given in the project website (https://sites.google.com/view/braidsideearth/). Current email contact of the Fellow is Gu.Stecca@niwa.co.nz.
The results achieved in different areas (geomorphology literature review, model development, model application) are complementary to each other and help unravel the mechanisms and dynamics of morphological change in impacted braided rivers. The model, trained on laboratory data and/or field data (when available) can be used to predict river behaviour in a countless number of different situations by simply changing inputs and controls. This helps scientists stretch their prediction ability well beyond the single cases used to train models, as our scenario applications show. These results are making in-roads towards the development of geomorphology as a quantitative and predictive discipline, and will potentially help devising mitigation and restoration strategies for impacted river systems.
Natural end users of our work are river managers, professionals, administrations interested in understanding the river evolution up to the present and to assess the impact of planned (changes in) management rules in the future. The numerical model developed within the project, despite being a research tool, represents the prototype for tools which we foresee will be of wider use among the above categories of stakeholders in the medium term (from a few years to a couple of decades). Before such a model can become a standard tool, however, some difficulties must be overcome. These difficulties are related to the high-demanding nature of research model both in terms of data and parameters to inform correct functioning of all the submodels (for bank erosion, vegetation) and to the very high demand of computational power needed to run long-term simulation in domains covering reasonably long and wide river reaches. Both these kinds of difficulties are foreseen to be progressively overcome in time with better spread and testing of these tools (which will help better understand how to correctly inform these with parameters), progressively improved data availability, and ever-increasing available computational power. In the meantime, there is room for a huge number of applied projects aiming to let the proposed model improvements leak into the most known and commonly used modelling frameworks, and to apply these tools to more test and pilot cases.