Cathy Kurz-Besson

The Water Isotopic Version of the Land-Surface Model ORCHIDEE: Implementation, Evaluation, Sensitivity to Hydrological Parameters

Land-Surface Models (LSMs) exhibit large spread and uncertainties in the way they partition precipitation into surface runoff, drainage, transpiration and bare soil evaporation. To explore to what extent water isotope measurements could help evaluate the simulation of the soil water budget in LSMs, water stable isotopes have been implemented in the ORCHIDEE (ORganizing Carbon and Hydrology In Dynamic EcosystEms: the land-surface model) LSM. This article presents this implementation and the evaluation of simulations both in a stand-alone mode and coupled with an atmospheric general circulation model. ORCHIDEE simulates reasonably well the isotopic composition of soil, stem and leaf water compared to local observations at ten measurement sites. When coupled to LMDZ (Laboratoire de Meteorologie Dynamique-Zoom: the atmospheric model), it simulates well the isotopic composition of precipitation and river water compared to global observations. Sensitivity tests to LSM (Land-Surface Model) parameters are performed to identify processes whose representation by LSMs could be better evaluated using water isotopic measurements. We find that measured vertical variations in soil water isotopes could help evaluate the representation of infiltration pathways by multi-layer soil models. Measured water isotopes in rivers could help calibrate the partitioning of total runoff into surface runoff and drainage and the residence time scales in underground reservoirs. Finally, co-located isotope measurements in precipitation, vapor and soil water could help estimate the partitioning of infiltrating precipitation into bare soil evaporation.

Effects of Recent Minimum Temperature and Water Deficit Increases on Pinus pinaster Radial Growth and Wood Density in Southern Portugal

Western Iberia has recently shown increasing frequency of drought conditions coupled with heatwave events, leading to exacerbated limiting climatic conditions for plant growth. It is not clear to what extent wood growth and density of agroforestry species have suffered from such changes or recent extreme climate events. To address this question, tree-ring width and density chronologies were built for a Pinus pinaster stand in southern Portugal and correlated with climate variables, including the minimum, mean and maximum temperatures and the number of cold days. Monthly and maximum daily precipitations were also analyzed as well as dry spells. The drought effect was assessed using the standardized precipitation-evapotranspiration (SPEI) multi-scalar drought index, between 1 to 24-months. The climate-growth/density relationships were evaluated for the period 1958-2011. We show that both wood radial growth and density highly benefit from the strong decay of cold days and the increase of minimum temperature. Yet the benefits are hindered by long-term water deficit, which results in different levels of impact on wood radial growth and density. Despite of the intensification of long-term water deficit, tree-ring width appears to benefit from the minimum temperature increase, whereas the effects of long-term droughts significantly prevail on tree-ring density. Our results further highlight the dependency of the species on deep water sources after the juvenile stage. The impact of climate changes on long-term droughts and their repercussion on the shallow groundwater table and P. pinaster’s vulnerability are also discussed. This work provides relevant information for forest management in the semi-arid area of the Alentejo region of Portugal. It should ease the elaboration of mitigation strategies to assure P. pinaster’s production capacity and quality in response to more arid conditions in the near future in the region.

Searching for an indicator of N evolution during organic matter decomposition based on amino acids behaviour : a study on litter layers of pine forests

Soil microflora can directly take up the amino acids (Aas) released by decomposing plants and use them, together with de novo synthesized Aas, as building blocks for their own structures, which are expected to have an Aa composition that differs from plant-derived structures. The degree of change in the Aa composition during OM decomposition should reflect the degree of N evolution. Here we apply this concept to the study of three litter decomposition sequences in European forests: boreal (Pinus sylvestris L., Jädraås, Sweden), cool Atlantic (Pinus sylvestris L., La Gileppe, Belgium), and warm Atlantic (Pinus pinaster L., Montemor, Portugal). Litter was sorted into six decomposition stages according to morphological features: from intact, light-brown needles to dark needle fragments with faunal perforations. At each stage, the Aa composition of the litter was studied by acid hydrolysis plus liquid chromatography of the hydrolysates. Both the Aa content of the litter and the unhydrolyzable/total OC ratio increased with decomposition, but this was not the case with the unhydrolyzable/total N ratio. There is no sole pattern of Aa change with decomposition: in Jädraås and Montemor most of the changes occur in the initial and final steps, while in La Gileppe the changes seem quite evenly spread throughout the decomposition. The fact that the Aa composition of the litter did not converge with decomposition suggests the existence of site-specific biotas with contrasted Aa fingerprints. The abundance of a given Aa relative to the total differed among Aas and sites; but for some Aas it was possible to detect common patterns of decomposition behaviour (either a consistent increase or a consistent decrease). The aa/total ratio for Gly, Ala and Thr consistently increased with decomposition. These Aas were used to search for a numerical index, the Aa signature (AaSIG), which would reflect the changes in Aa composition from fresh debris to highly decomposed organic materials. The change in the proposed AaSIG was found to correlate well with the mathematical distance from the starting point (calculated as the euclidean distance); this suggests that it could be potentially useful as an indicator of N evolution during litter decomposition, at least in its first stages.

The propagation of particulate gravity currents in a V-shaped triangular cross section channel: Lock-release experiments and shallow-water numerical simulations

In this study, we investigate the motion of particulate gravity currents in a horizontal V-shaped channel. The particulate currents consisted of particles whose size varied between 0 and 100 μm but whose mean size increased. Particles were poorly sorted as the variance of the grain size distributions varied between 50 and 200. While the phases of propagation of homogeneous currents in such a geometry have been studied in the literature, this study considers the effects of the grain size on the propagation. The distance of propagation and front velocity of full-depth high-Reynolds-number lock-release experiments and shallow-water equation simulations were analyzed as the mean grain size of the initial particle distributions, defined by mass, was increased from 19 to 58 μm. Similar to the homogeneous currents, three consecutive phases of the front velocity could be identified but their characteristics and extent depend on the particle size. The initial phase, in particular, depends on a dimensionless settling number β that is defined as the ratio of two characteristic time scales, the propagation time x0/U, where U is the scale for the front speed and x0 the lock length, and the settling time h0/vs, where vs is the scale for the settling velocity and h0 the initial height of the current. For dimensionless settling numbers less than 0.001, the initial phase is characterized by a constant velocity for over about 6-7 lock lengths that is alike the initial slumping phase of perfectly constant velocity of the homogeneous currents. For dimensionless settling numbers greater than 0.001 and less than 0.015, the initial phase is no longer characterized by a constant velocity but an almost constant velocity for over about a similar 6-7 lock lengths. For dimensionless settling numbers greater than 0.015, however, as such, this phase is no longer seen. This initial phase is followed by a continuous decrease of the front advance, which results from the sedimentation of the particles. Unlike the homogeneous currents, this phase is a non-self-similar propagation. This phase is ended by a viscosity-dominated phase appearing to vary as ∼t1/7. The good agreement between the front advance of the experiments and shallow-water equation simulations demonstrates that the mean size by mass is a fairly good proxy of poorly sorted particles.