ALBERT

Beschreibung: Knowledge of soil moisture dynamics and its spatial variability is essential to improve our understanding of root water uptake and soil moisture redistribution at the local scale and the field scale. We investigated the potential and limitations of electrical resistivity tomography (ERT) to measure three-dimensional soil moisture changes and variability in a large, undisturbed, cropped soil column and examined the interactions between soil and root system. Our analysis sustained the value of ERT as a tool to monitor and quantify water contents and water content changes in the soil, as long as the root biomass does not influence the observed resistivity. This is shown using a global water mass balance and a local validation using time domain reflectometry (TDR) probes. The observed soil moisture variability was rather high compared to values reported in the literature for bare soil. The measured water depletion rate, being the result of combined effects of root water uptake and soil water redistribution, was compared with the evaporative demand and root length densities. We observed a gradual downward movement of the maximum water depletion rate combined with periods of redistribution when there was less transpiration. Finally, the maximum root length density was observed at -70 cm depth, pointing out that root architecture can strongly depend on soil characteristics and states.

Beschreibung: A lot of data is available in the literature on root length density profiles (i.e., root length per unit of soil volume versus soil depth), because they are a traditional way of representing root distribution in the field. In a complementary approach comprehensive models of the root system architecture are being developed, but the parameters required for these models are often difficult to assess in field conditions. In this paper, we bridge both approaches, empirical and comprehensive, by evaluating the capability of architectural models to simulate the observed diversity in root length distribution on the one hand and the possibility of estimating developmental parameters from root length density profiles on the other. For this purpose, we constructed a simple model with only six parameters, to represent the root system architecture. It reproduced a large diversity of root profiles comparable to observations reported in the literature and encompassing many different crops. The impact of each model parameter, as well as their interactions, on the shape of the profiles was quantified using a global sensitivity analysis. Finally, a statistical meta-model was designed and estimated to simulate the same collection of profiles without intermediate simulating the whole architecture. The meta-model allows for estimation of architectural parameters from profile shapes (inversion). Some architectural parameters could be estimated from the profiles with good accuracy, especially those quantifying the growth potential and gravitropism of individual roots, because of their specific impact on root length at a specific depth. But others (like inter-branch distance, life duration), which modify root density in a more diffuse way throughout the profile, could not be identified correctly using this method. Additional data involving specific measurements are necessary to identify these last parameters.

Beschreibung: The development of models describing water and nutrient fluxes to and through 3-D spatially resolved root structures in soils brings along the need to predict or describe the root architecture and root growth in detail. However, detailed data to calibrate and validate such architecture and growth models is typically not available. Here, we investigate the sensitivity of the root architecture model RootTyp to changes in its model parameters and reconstructed the root system architecture of barley ( Hordeum vulgare L.) growing in an undisturbed lysimeter using minirhizotron images at four different depths. Root arrival curves from a series of minirhizotron images were used to parameterize RootTyp using a range of realistic architectures. We adjusted a simple architecture to the data, which contained only long primary roots starting from the seed. This simple model unfortunately could not reproduce the observed increase of root density with depth. The model was subsequently improved by allowing root branching and elongation to be horizon-dependent and by making reiteration of root tips possible. Reiteration is an alternative form of branching, where secondary roots can become as long and thick as primary roots. Our results show that minirhizotron data do not contain enough information to warrant identification of the parameters governing these processes, as the additional parameters act similarly on data characteristics as the initial ones. Therefore, different experimental techniques should be combined to constrain the model parameters better in the future.