Periodic Reporting for period 1 - MOSS (Models of Soil Hydraulic Properties with Adjustable Soil Structure and their Application in an Earth-system Model)
Période du rapport: 2023-10-01 au 2025-09-30
Computer models used for predicting the weather or the climate must calculate how much water from rain and snow enters the soil and for how long the water stays close to the surface, where it can evaporate (and thus cool the land surface) or be taken up by plants. To calculate this, land components of weather and climate models require parameters which characterize the soil at a given location. In most land models, these parameters are estimated based on the proportions of small, medium, and large particles that make up the soil, without taking into account how these particles are arranged. Not accounting for soil particle arrangement -- the soil structure -- means that land surface models cannot account for many processes which are known to affect soil parameters, such as biological activity (e.g. earthworms), land management (e.g. tillage, compaction), or wet-dry and freeze-thaw cycles. This potentially makes weather and climate predictions less accurate, and limits the extent to which land models can weigh in on timely issues involving soil biodiversity preservation or agricultural practices with strong impact on the soil. The goal of MOSS was to develop a flexible, physics-based method for accounting for the effects of soil structure on soil parameters, implement the method in a state-of-the-art land surface model, and investigate the impact of soil structure on soil moisture and other land variables.
The work performed in MOSS is divided into three packages: the development of the MOSS method, its implementation in the ORCHIDEE land surface model, and simulation experiments with the resulting ORCHIDEE version. The main activities in the development package included 1) a through literature search, which allowed us to connect the MOSS method to related developments and to potential applications; 2) the development of a reference Python implementation of the method, which has been open-sourced; 3) the application of the MOSS method to site-level case studies involving soil recovery following compaction by heavy machinery and soil structure formation as a result of long-term organic amendment. Following the development of the method and the preparation and submission of a manuscript detailing the process and the results, the method was implemented in the ORCHIDEE model, which is the land surface component of the IPSL-CM Earth system model. This work package included 1) Adaptation of the MOSS method to a saturation-based (as opposed to pressure-based) form of Richards's equation used in ORCHIDEE; 2) Implementation of the MOSS method in Fortran (the programming language of ORCHIDEE); 3) Testing of the ORCHIDEE Fortran implementation of the MOSS method against the reference Python implementation. Following the implementation of the MOSS method in ORCHIDEE, we performed simulation experiments investigating the large-scale effects of soil structure on the terrestrial water balance, and the extent to which they are mediated by model subgrid parametrizations. The main activities in this work package included 1) The design of the simulation experiments (a land-only factorial experiment and two long coupled land-atmosphere runs); 2) Running the simulations; 3) Statistical analysis and visualisation of the data. Based on the results of this work package, a second manuscript was prepared and submitted.
This project formulated and disseminated a unique method of incorporating multidimensional (beyond bulk density) soil structure data into ecohydrological and land surface models. Both a theoretical description and a reference Python implementation of the method were published and are freely available, encouraging reuse and exploitation. The method was also implemented in a state-of-the-art land surface model (ORCHIDEE), demonstrating the feasibility of accounting for and the impacts of soil structure at large scales. These impacts include significant regional cooling, the decrease in surface runoff, as well as increases in deep drainage, total soil moisture, and transpiration as a result of accounting for plant-linked soil structure. Investigating the role of subgrid parametrizations in mediating the large-scale effects of soil structure clarified the prerequisites necessary for the representation of soil structure in land surface models. These include representing, to some extent, the subgrid variabilities of precipitation or of soil hydraulic properties. Most elements necessary to ensure further uptake and impact of the results are in place. The impact could be further amplified though follow-up research involving the inverse modeling of soil structure based on remote-sensing observations.