The meteorological development of 2018 has caused droughts followed by high losses in yields and yield quality among most agricultural crops for wide parts of Germany. Considering future trends, as featured in the latest IPCC report, irrigation will play a key role, also in predominantly humid environments such as Germany, to ensure farmers prosperity whilst maintaining high levels of food security.
However, sustainable forms of irrigation, automated or not, require reliable input data on biomass status and other environmental variables such as soil moisture. Considering the latter, this paper introduces a recently explored technique for soil moisture retrieval based on GNSS-data, which is intended to provide both, a reliable source for in-situ validation of remotely sensed data, as well as a low-cost, non-destructive and easy-to-operate system for soil moisture measurements at industrial scale farms.
Based on studies by Koch et al. (2016), the freely available L-band signal of GNSS systems (currently Galileo and GPS) is used in a complex retrieval scheme to determine soil moisture of bare fields by analyzing the signal attenuation that occurs between sensors on the soil surface and sensors buried in the soil profile. In addition, the effects of dynamic biomass evolving above the ground-based sensors is studied. The station´s hardware is based on a modified version from the ESA business applications co-sponsored demo project SnowSense (2015-2018). Featuring autonomous power supply and mobile network communications for fast data processing and remote maintenance, the station uses 3 GNSS receivers (one on a 2 m mast for reference and baseline calculations, one on the soil surface below biomass and one at 5 cm soil depth) to calculate respective differences in signal strength and possible attenuation by growing biomass. A first fully operational station has now been operating since September 2018, as well a sibling station since early October 2018, on mustard fields near Straßkirchen, approx. 130 km NE of Munich, Southern Germany. Datasets from several beta-stations are available throughout 2018 for different crops. Following installations, an extensive field campaign was carried out to obtain validation datasets on soil moisture, biomass and canopy water content. To ensure low costs for production and maintenance, consumer-grade GNSS receivers are used. These reveal sensor specific receptivity pattern for both dimensions (x and y), which is why spatial and temporal averaging is performed. Also, as the scattering characteristics change with different elevation angles, elevation classes are defined.
Overall, two different scenarios are covered: I) measurements with no biomass. Results obtained using this method agree with validation data, as was also stated by Koch et al. (2016). II) measurements below biomass. A considerable bias is introduced to the measurements by biomass and it´s respective dynamics. Vegetation scattering and attenuation effects are mainly influenced by the geometric structure and dielectric properties. Due to the high complexity of modeling microwave-plant-interactions and the lack of coherent data (phase information unusable, no polarimetric information available), signal attenuation was highly simplified.
For a first analysis, a basic two-box-approach was chosen. One box represents the soil layer, assuming homogeneity in structure and moisture within the first 5 cm. The second box abstracts the attenuation and scattering processes of the vegetation layer along the entirety of its canopy height. An adapted version of the water cloud model by Attema and Ulaby (1977) is used to simulate the attenuation process in dependence of vegetation properties. Since the station has only been in operation since few weeks, no final results can yet be shown here. However, first results on biomass effects on the quality of the GNSS-based soil moisture retrieval will be presented at the symposium in May 2019.

references:

Koch, F., Schlenz, F., Prasch, M., Appel, F., Ruf, T., Mauser, W. (2016): Soil Moisture Retrieval Based on GPS Signal Strength Attenuation. In: Water, 2016(8, 276), 1 – 22.

Attema, E., Ulaby, F. (1977): Vegetation modeled as a water cloud. In: Radio Science, Vol. 13(2), 357 – 364.

Area A - Board 148

In the context of global warming and the need to improve predictions of the climate trajectory of the 21st century, it is fundamental to be able to accurately estimate and understand trends of key variable as the surface soil moisture (SSM). This latter, recognized as "essential climate variables" (ECVs) by the Global Climate Observing System (GCOS) in the context of the United Nations Framework Convention on Climate Change (UNFCCC), has a large impact on water and energy fluxes, on the heterotrophic soil respiration and control the feedback between soil, vegetation and atmosphere. In this frame the Satellite Application Facility on Climate Monitoring (CM SAF) develops archives of satellite-derived products to support the understanding the climate. For instance, the CM SAF will extend its product portfolio with a Thematic Climate Data Record (TCDR) of sensible (H) and latent (LE) heat fluxes and evapotranspiration (ET). The retrievals are based on the methodology developed by the Royal Meteorological Institute of Belgium (RMI) for the LSA SAF (SAF on Land Surface Analysis; Ghilain et al., 2011, 2017) where the SSM is a key input. Thus, it is crucial to get homogeneous long term time series (at less few decades) and catch both long and natural or human induced short-scale trends of SSM. Furthermore, the European Spatial Agency (ESA) Climate Change Initiative (CCI) created a 38-year (November 1978 - December 2016) SSM climate data records by blending satellite measurements from active and passive microwave sensors. Recently, Ghilain et al., 2016 developed a new approach to derive the SSM from the land surface temperature data from thermal infrared sensors aboard geostationary satellites (referred as LST-SSM henceforth). Soil moisture trends can also be obtained from the ECMWF ERA5 modeled surface soil moisture at global scale. This type of dataset can induce a dependency when using it as input of models. Each product has distinctive limits which make challenging to derive accurately the SSM from unique approach. The objective of this study is to take advantage of each approach by using the more appropriate dataset according to the land cover while maximizing the use of satellite data to avoid potential model dependencies. First, the ESA-CCI SSM, the LST-SSM and the ERA5 rescaled (between 0 and 1) products have been evaluated by comparison with in situ datasets, provided by the International SSM Network and FLUXNET, with a large regional coverage and spanning multiple seasons which include measurements from 52 stations. Best result were obtained for the ERA5 dataset with a R2, a bias and a mean absolute difference (MAD) respectively of 0.66, -0.05 and 0.19, while they were respectively of 0.49, -0.05 and 0.21 for ESA-CCI SSM and 0.35, -0.005 and 0.23 for LST-SSM. Second, the stations have been sub-divided in three groups using a k-means approach based on the Leaf Area Index (LAI; GLOBal MAP dataset; Liu et al., 2012) informations and the climatological classes of the station (adapted from the Koeppen-Geiger classification). Finally, each group was associated to the product where the closest values with in-situ data were obtained and to specific ranges of LAI values (used as criteria selection). By applying the LAI criteria, the ERA5, the ESA-CCI SSM and the LST-SSM products were used respectively for 32, 17 and 3 stations. The R2, the bias and the MAD became of 0.6, 0.009 and 0.18. By applying the LAI criteria to the global map, it resulted that the ERA5, the ESA-CCI SSM and the LST-SSM products were obtained for respectively 60%, 15% and 25% of the total land surface. The ESA-CCI SSM and the LST-SSM products (remote sensing datasets) were mainly selected in semi-arid and boreal regions.

Area A - Board 149

The largest European rice production is located in northern Italy, where paddy field flooding management shifted towards a more environmentally friendly and saving practice (dry seeding) that has already reached already its sustainable potential. This resulted in the largest anthropic reduction of seasonal surface water extent detected in Europe after 2000, concurrently with a climatic shift that contributed to higher rice yields and stability. Before 2000, rice fields were always flooded and yields were strongly anti-correlated to the downstream river discharge, which acted as a proxy for unfavorable weather conditions. Analogue conditions currently characterise other important rice producing countries such as Uruguay, Argentina, North and USA. In the context of near-term climate changes, sustainable rice cultivation in the middle latitudes seems therefore achievable without limiting production and/or increasing volatility. We discuss the environmental pros and cons as well as the specific requirement needed by a sustainable shift to rice dry seeding.

Area A - Board 150

Earth Observation climate data archives reach now a substantial long-term period that allow potential studies for monitoring the mechanisms underlying the climate change dynamics. The Copernicus Climate Change service (C3S) through its Climate Data Store (CDS) provides an easy access to several long-term climate data. It is therefore important to assess the agreement between model and EO climate products before using them in climate change dynamics.
The aim of this presentation is to infer the coherency of various long-term climate variables by analysing their trends over terrestrial regions. As example, trend of surface albedo and land surface temperature (LST) derived from difference sources of earth observation and numerical models will be shown.

Area A - Board 151

Long (more than 30 years) time series of satellite-derived products over land are now available. They concern Essential Climate Variables (ECV) such as LAI, FAPAR, surface albedo, and soil moisture. The direct validation of such Climate Data Records (CDR) is not easy, as in situ observations are limited in space and time. In the framework of land reanalyses, the consistency of terrestrial ECVs must be ensured. The most advanced cross-cutting quality monitoring technique consists in integrating satellite-derived products into a land surface model (LSM) through a land data assimilation system (LDAS). The reanalysis produced by the LDAS accounts for the synergies of the various upstream products and provides statistics, which can be used to monitor the quality of the assimilated observations (Zeng et al., https://doi.org/10.1016/j.jag.2015.06.006). CNRM has developed the open-source offline LDAS-Monde chain (Albergel et al., https://doi.org/10.5194/gmd-10-3889-2017), based on the ISBA LSM. The LSM is able to represent the diurnal cycle of the surface fluxes, snow, root-zone soil moisture profile and soil temperature profile, together with the day-to-day, seasonal, interannual and decadal variability of LAI and vegetation biomass. The LSM is embedded in the SURFEX modeling platform together with a simplified extended Kalman filter and an Ensemble Kalman filter. It is coupled to the CTRIP river discharge model (Decharme et al., https://doi.org/10.1007/s00382-011-1054-9). The current version of LDAS-Monde is able to assimilate LAI (or FAPAR) and surface soil moisture (SSM) products at a spatial resolution of 0.25 degree x 0.25 degree at a global scale, and at enhanced spatial resolutions (as high as 1 km x 1 km) over any small region. A passive monitoring of albedo, FAPAR and Land Surface temperature (LST) is performed (i.e., the impact on these variables of assimilating LAI and/or SSM is assessed). The LDAS-Monde system is used in the European Copernicus Global Land Service (http://land.copernicus.eu/global/) to monitor the quality of upstream products at a global scale. Scores are given for 19 key areas which experienced severe droughts or heat waves in the past (e.g. western Europe, western Russia, India, southwestern China, US southern plains, Brazilian Nordeste, Australian Murray-Darling basin, Sahel, ...). Results are presented for the 2008-2018 time period. Plans to use LDAS-Monde in the CCI+ Climate Modelling User Group (CMUG) are presented. In this framework, LDAS-Monde will be used to assess the consistency of snow, LST and permafrost with other ECVs. The assimilation of snow fractional coverage will be implemented, together with the passive monitoring of snow water equivalent.

Area A - Board 152

The management of water resources in agricultural areas requires a precise knowledge of the evapotranspiration (ET) which is largely investigated over plan region. Whereas, the evapotranspiration over mountain regions still currently not very known due to the complexity of the terrain. In general, the complexity is associated with the type and the cover of the vegetation canopy as well as with changes in topography, which make the heat and mass transfers more complicated than over homogenous and flat surfaces. In the present study, the issue of using a Two-Source-Energy-balance model (TSEB) driven by the LANDSAT data for estimating evapotranspiration over Tahnaout site has been investigated. This site, which located about 35 km to Marrakech city (Morocco), consisted of a mixed of arboriculture, wheat, corn, alfalfa, follow and bare sol. Therefore, the heterogeneity caused either by the contrast in soil moisture, cover or slope generates a large variability of fluxes which limit the use of local scale measurements device such as the eddy covariance system (EC), which considered as the only method that can measure directly ET with high precision, unless one deploys a network of EC devices which is costly and requires a continuous availability of well-trained staff to operate and maintain the devices. In this context, a number of different techniques have been introduced for research applications, such as displaced-beam laser scintillometers, and large or extra-large aperture scintillometers (LAS, XLAS). Compared to eddy covariance systems, the scintillometer is easy to install, relatively cheap and it is a practical method to obtain area-average surface fluxes over several kilometres. The instrument is capable of continuous measurements with minimum human intervention. However, this instrument measures only the sensible heat fluxes and the evapotranspiration can be derived as the residual term of the energy balance equation. In this study, a Large Aperture scintillometer (LAS) was installed over a transect of about 1.4 km to measure the convective fluxes. The LANDSAT-7 and LANDSAT-8 data used as inputs for the TSEB model such surface temperature, albedo emissivity and Leaf Area Index (LAI) derived from NDVI, were firstly aggregated based on the values of LAS footprint. The results obtained by comparing the evapotranspiration values simulated by the TSEB model and measured by the LAS showed good consistency (R = 0.72, RMSE = 52.4 Wm-2, bias = 19 Wm-2). On the other hand, TSEB slightly underestimates the sensible heat flux but the comparison remains acceptable (R = 0.65, RMSE = 75.5 Wm-2, bias = 56.2 Wm-2).

Area A - Board 153

A large part of the Earth's surface has been modified for urban and industrial development, as well as for agricultural and livestock farming purposes. Almost 35% of the land surface has undergone changes as a direct consequence of human interaction with the natural landscape. Until the mid-20th century most land-use changes took place over the temperate regions of the Northern Hemisphere. Along with changes in terrestrial ecosystems, a change in land-use has impacts on climate through modification of the surface energy and water budget, mediated by changes in biophysical effects like surface albedo, evapotranspiration, and surface roughness. However, the relative biophysical effects of recent land-use changes, mainly in the mid-latitudes and the European region at a regional scale have not yet been quantified. The regional climate community can strongly benefit from the state-of-the-art satellite observations, which now provide high resolution observations of land-use change from space. The newly released annual ESA CCI (Climate Change Initiative) land cover maps provide long-term time series of land-cover change at 300m resolution, and can be used in land surface models (LSMs) and climate models to simulate LULCC effects on surface energy budgets. The innovation we propose in this work is to incorporate the high-resolution state-of-the-art ESA CCI land cover product into regional climate models (RCMs), in order to estimate the magnitude and signal of climate change in finer spatial scales over Europe, including the impact of land-use change. This work is part of the international initiative LUCAS, which is an endorsed Flagship Pilot Study of the regional climate CORDEX community, operating under the auspices of the World Climate Research Program (WCRP).

Area B - Board 154

There is constant demand from industry for high-quality, long-term gridded datasets with high temporal and spatial resolution for conducting climate research. These data have potential to be used in fields such as agriculture, water resource estimation and management, hydrology and hydrogeology, public health, energy and planning and studies on observed climate change trends and vulnerability. Variables such as evapotranspiration and soil moisture conditions are crucial factors in estimating water sustainability, understanding groundwater recharge, agronomic management and the management of flood and drought risk. However, with the exception of temperature and precipitation, spatially and temporally homogeneous, multi-decadal, gridded observational climate datasets are not readily available for hydro-climatic research applications in Ireland.

One approach to generating the requisite gridded datasets is through the use of NWP downscaled simulations and can also be used to “reanalyse” historical weather conditions, whereby past weather observations such as satellite, surface and upper-air conditions are used as model inputs. This report describes the application of downscaled NWP models to produce gridded climate datasets of key hydro-climate variables for Ireland.

In this research the performances of three NWP models [COSMO-CLM, WRF and Met Éireann Re-Analysis (MÉRA)] are compared and analysed to assess their ability to accurately represent hydro-climate variables, focusing on reference evapotranspiration as calculated using the FAO Penman–Monteith equation. Two of the NWP datasets assessed were produced by researchers at the Irish Centre for High-End Computing (ICHEC) using the COSMO-CLM5 and WRF v3.7.1 models. The third dataset (MÉRA) is derived from a 36-year (1981–2016) regional reanalysis of the Irish climate carried out by Met Éireann using the HARMONIE model and the ALADIN-HIRLAM NWP system. The MÉRA simulation included a data assimilation component and has a resolution of 2.5km. All NWP models used in this study were downscaled using the European Centre for Medium-Range Weather Forecasts (ECMWF) ERA-Interim global reanalyses dataset.

Modelled reference evapotranspiration (ET₀) data are analysed and validated against ET₀ data calculated from meteorological observations at 22 Met Éireann synoptic stations across Ireland. Least squared estimator monthly correction factors were applied to all three raw datasets to improve the output. Using these ET₀ datasets, actual evapotranspiration (ETa) and soil moisture deficits (SMDs) have been derived using the hybrid soil moisture deficit model, created by Teagasc and in use in Ireland since 2006. The following datasets were derived using MÉRA input variables containing daily, monthly, seasonal and yearly time steps:

• ET₀;
• ETa (five drainage classes);
• SMDs (five drainage classes).

The datasets described are also available for the COSMO-CLM and WRF models with 1.5-km and 2-km grid spacing, respectively. Additionally, the Teagasc Indicative Soil Drainage Map has been implemented to capture the actual soil conditions throughout the country. All datasets are available for download through the ICHEC ERDDAP server.

Across all three models, the MÉRA model is the best performer for reference evapotranspiration. Using a weighted average across 22 synoptic stations, root mean square errors (RMSEs) equated to 0.337 mm/day for the MÉRA model compared with 0.402 mm/day and 0.442 mm/day for the WRF and COSMO-CLM models, respectively. The MÉRA model also performs best for the ETa and SMD variables and therefore the MÉRA datasets are the recommended datasets to be used as indicative maps for Ireland.

Area B - Board 155