Fluvial Vegetation Dynamics: An Innovative Approach to Modeling Root Biomass in Riparian Environments

The aim of this thesis work is to investigate on the variety of the most common modeling approaches used to describe vegetation dynamics in river systems and to propose improvements to some of their key aspects. According to a large body of literatures the multitude of external factors that influences vegetation evolution – both positively and negatively - makes it difficult to construct a complete and realistic mathematical model. For this reason, most studies focus only on the most significant factors, such as groundwater level, which has been widely shown to strongly affect vegetation behavior. A significant portion of the thesis is dedicated to a critical review of some of the most well-known vegetation modeling approaches. This analysis aimed to identify which aspects are consistently addressed and which are still treated with insufficient detail. Among these, root biomass emerged as a crucial yet oversimplified component despite its role in the interaction between vegetation and hydrological conditions, Therefore ,this works focuses mostly on the role that the root system plays in the modelling approach and how it is influenced by groundwater level variability, in order to prioritize root biomass as the main characterizing variable of the dynamic computation. Starting from a pre-existing root biomass model that describes dynamics of phreatophytic species, typically found in riparian environment, a new implementation is proposed in the form of a 2D time depending computation. This model simulates local variation of the biomass in the subsoil and of the rooting depth, based on groundwater level, distance from the riverbed and vegetation characteristics. The model has been tested under different conditions in order to assess its validity and potential. Different geometries were considered – both linear and irregular – and the input hydrological regime was implemented in both statistical (sinusoidal) and stochastic form. In the latter case, a water level generator was used to reproduce the natural variability of river flow, by imposing different conditions such as varying coefficients of variation and integral time scales. Ultimately, statistical analysis over the results of the simulation has been conducted to identify the most influential factors governing the system, with the aim of improving the predictive capabilities of vegetation dynamics models in riparian environments.