Probabilistic short-term to seasonal soil moisture forecasts over Germany and surrounds using the hydrologic model ParFlow/CLM

Abstract

<p><span>The repeated droughts that have affected Europe in 2018, 2019, and 2020, as well as increasing environmental and legal constraints that demand for a more sustainable management of soil water resources, stress the needs for terrestrial water budget forecasts at stakeholder-relevant high spatial resolutions, e.g., for the agricultural sector. In this context, we have established and operate a new quasi-operational forecasting experiment with the integrated hydrologic model ParFlow (www.parflow.org) which forecasts all relevant states and fluxes of the terrestrial water cycle over Germany and its neighbouring regions at 600m horizontal resolution. ParFlow simulates the complete 2D/3D surface and subsurface water budget, including the unsaturated and the saturated (groundwater) zones, as well as the energy and water fluxes at the atmosphere-surface-subsurface interface through its surface module CLM (Common Land Model). The simulations are driven by surface and near-surface atmospheric variables from different weather forecasting products from ECMWF (European Centre for Medium-Range Weather Forecasts). The 10-day hydrologic forecast, driven by the deterministic high-resolution weather forecast HRES, is complemented by a 50-member ensemble of ParFlow/CLM simulations forced by ECMWF’s ENS probabilistic forecast ensemble. This short-term ensemble approach allows for evaluating the impact of the uncertainties of the weather forecast, and especially of precipitation, on the different components of the terrestrial water budget. At the beginning of each season, we additionally run a 4-month seasonal prediction ensemble of the terrestrial water cycle using ECMWF’s SEAS long-term probabilistic 50-member ensemble forecast. Since soil moisture has a high inertia, especially for deeper soil layers, compared to, e.g., atmospheric variables, the spread of the ensemble remains acceptable over the prediction period. To provide the most useful information products on the evolution of water resources over the upcoming months to stakeholders, we derive diagnostics from the probabilistic seasonal forecasts that allow for a risk assessment. For example, for each 10-day period of the forecast, we calculate the probability (or risk) for the plant available water to reach values below a critical threshold for water stress, e.g., 30%. A prototypical application of these experiments is the ADAPTER project (www.adapter-projekt.de). In this framework, the monitoring and forecasting of relevant soil moisture-related diagnostics is meant to provide support for agricultural and water resources management, e.g., against droughts.</span></p>